You tell yourself the company will become easier to hold together once there are other people helping. It should stop depending so much on you and your ability to keep dragging things forward.

Across Millennial Masters interviews, I’ve spoken to founders who got there, with hundreds of staff and revenues running into the hundreds of millions, and sometimes beyond. Then the worries changed.

The old business started failing before the new one was fully built. The job changed, and they moved further away from the work that made building it exciting in the first place. 👇🏻

The old company breaks first

Joshua Western said going from 50 people to 60 was harder than going from 2 to 50. Around 50 people, “the things that you set up, the ways of working, what little processes you try and have as a startup, all fall over.”

Growth puts more pressure on the company and shows how much of it was being held together by people who knew what to do without being told.

A lot of what looked solid was really people filling gaps, remembering what mattered, and sorting things out because the team was still small enough for that to work.

Then the company reaches a size where those old shortcuts stop working, and suddenly simple jobs need more chasing than they used to.

Emma Mills described this break point. Once a company reaches seven figures, what got it there doesn’t get it to the next stage.

The early version relies on people doing whatever needs doing. The next one needs clearer roles and fewer important calls stuck in one person’s head.

That can seem daft when you built the business on speed and instinct. If you ignore it, the problems show up somewhere else.

Headcount can lie

Noel Andrews dealt with one of the easiest founder vanities: headcount.

He scaled to 17 people and thought that meant the business was moving forward. From the outside, it looked like a bigger, more serious business.

Inside, it was chaotic. The company looked like “a graceful swan on the surface,” while underneath there was too much paddling and too little control.

Then he went back down to eight people. The smaller team did twice as much work and made three times the profit.

More people can make the business look more real. They can also create work that only exists because more people are there.

Add people without rebuilding how the company works, and you can make the business look bigger from the outside while it gets worse inside.

The job changes on you

Tyler Dunagin looked back on his own journey and described the founder role as a move from all-star NBA player to coach, and later towards the owner’s box.

At the start, you’re on the field. You’re doing the work, solving all the problems. Then scale starts pulling you away from that closeness.

It sounds sensible until you’re the one having to make the swap. Some founders are good at the early fight and much less comfortable in the role the company needs later.

Nick Telson-Sillett recalled meeting people by the coffee machine in his own office and realising he didn’t know them because he hadn’t been involved in hiring them. They would ask what he did, and he would say he was the founder.

You can become a stranger inside the company you built. That’s the part founders don’t always see coming.

The job changes from being close to the work to building a company that can keep moving without you touching everything. Yet you can still miss the version of the business where you felt useful in a more immediate way.

Millennial Masters

Business, growth, and AI. The newsletter and podcast for founders and leaders.

By Daniel Ionescu

Early heroes can turn blockers

Simon Penson said it’s the growth stage where someone in the room feels so important that, if they left, the business would be in trouble.

Startups depend on people like that. They stay late and help the company survive when there’s no real structure around them.

Then the business grows, and the person who was essential at the start can become a blocker in the next stage, because that step needs something different.

A practical generalist can struggle once the business needs a specialist. The person who did three jobs may not be the right person to lead one of them at scale.

That’s hard because the role has history attached to it. You remember who helped when the business was still fragile and the future wasn’t always clear.

Their contribution can still matter. The problem starts when you keep them in the same role because you feel you owe them.

Then everyone else has to work around it. People cover the gaps, standards slip, and the business starts fitting around the person instead of the work.

That’s how loyalty becomes expensive.

Clarity becomes part of the work

In the early stage, people sit near each other and things keep moving because whoever needs to fix the problem is close by. You can get away with a lot when everyone knows the backstory.

Nick Holzherr remembers the gap between 30 people and 120 people. At that size, you have to think much harder about how things are said and done, so they still make sense when you’re not there for every decision.

More time goes into writing things down so people can act without you explaining it each time. It’s necessary work, and it’s less fun than a quick conversation with those who already understand the business.

Andrew Steele understood a business with 100 people shouldn’t feel like a startup with five. You can’t just run it over WhatsApp any more.

You can keep trying to hold on to the startup feeling long after you’ve outgrown it. Eventually, people need to know what good looks like without asking you every time.

That may feel slower, but it’s just what happens when the business is no longer small.

The numbers get heavier

From the outside, scale can make a company look safer. More revenue and a bigger team suggest the fragile early stage is over.

James Fleming looked back at the early years and said that if he’d seen the later outgoings, he would have cried. The business reached around £135,000 a month just to keep running.

That kind of number changes the company. You can have more revenue and feel less free because the cost of being wrong grows with the business.

Kevin de Patoul lived the hypergrowth version. His company went from 15 to 60 people, then after raising $70 million, from 60 to 160 in 12 months. His warning was that if you keep pushing forward without looking at what’s breaking behind you, things can start imploding.

That’s the danger of scale. The business can look stronger from the outside while small problems are getting more expensive inside it.

The job you wanted disappears

Joshua Dziabiak loved the stage where the job was still about making something. You could take an idea from nothing and turn it into a real business.

Then the company scaled, and the work changed. He described it as “financial engineering.” The big moves became smaller, more careful ones. The work became about keeping a larger machine on track.

The business had succeeded. The work that made it feel alive started to disappear.

Joshua also described the strain of carrying that stage. You put on a brave face, tell the story, get people fired up, then go to bed wondering whether you’re losing your mind. He joked that building the business took 10 years off his life.

You can be exhausted. The company can take too much from the person leading it. The role can become one you want, without wanting to perform it every day.

Joshua eventually recognised that he was a zero-to-100 founder. He wasn’t the person who wanted to take a company from $100 million to $10 billion.

That kind of self-knowledge is rare because founder culture rewards the opposite instinct. You keep going and keep telling yourself that stepping aside would look like defeat.

Sometimes the honest answer is that the company has outgrown the work you’re best at.

Scale is a different company

That doesn’t mean the founder failed. It means the company changed.

At some point, you have to ask whether you still want the job that comes with the next stage.

The old company may have to break for the next one to exist. The harder part is that it can break something in you too.

You can chase scale as the reward for building well, then realise it’s given you a company that needs a version of you that was never really yours.

That’s the cost you don’t always see until you’ve already built the thing you thought you wanted.

Millennial Masters

Business, growth, and AI. The newsletter and podcast for founders and leaders.

By Daniel Ionescu

More from this Millennial Masters series:

Your profile and posts are becoming part of how professional authority gets found and described beyond the platform.

, writer of The Link Tank, has spent more than 15 years helping professionals build credibility on LinkedIn. In this guest piece, she explains what you need to fix now. 👇🏻

AI search is reading your LinkedIn

Your next prospect may not Google you. They may ask an AI assistant, and LinkedIn is becoming one of the sources those systems lean on when answering professional queries.

Many founders still aren’t thinking about LinkedIn this way. But what they put there, across their profile and posts, is starting to do more than support human search.

It’s becoming part of the evidence layer AI tools use when they decide how to describe someone professionally.

The old rules of SEO and keyword-optimised profiles still matter. There’s now another layer on top.

The numbers have changed

LinkedIn is now the most-cited domain for professional queries across six major AI platforms. ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Microsoft Copilot, and Perplexity all cite LinkedIn more than any other professional source.

The growth has been fast. Between November 2025 and February 2026, LinkedIn moved from outside the top twenty to among the most-cited sources on ChatGPT alone, with citation frequency more than doubling in three months.

A Semrush study of 89,000 LinkedIn URLs cited in AI responses found that LinkedIn was the second-most-cited domain overall, ahead of Wikipedia, YouTube, and every major news publisher. Across AI responses, 11% referenced LinkedIn. On ChatGPT Search specifically, that figure was 14.3%.

Meanwhile, more buyers are using AI tools to research people, companies, and services before making decisions. Your profile is increasingly being evaluated by something before a human sees it.

AI is building a picture of you

AI tools don’t just send people towards you the way Google used to. They can describe you inside the answer itself.

When someone asks ChatGPT, “Who are the best advisers for fintech compliance?” or asks Perplexity, “Who should I follow for startup fundraising advice?”, those tools pull from public sources, including LinkedIn content and profile data, to construct an answer.

This is where the shift becomes more important for founders. AI tools are increasingly indexing and citing the content people create on LinkedIn, not just their profiles.

Profound’s longitudinal data shows that citations to LinkedIn feed posts and long-form articles grew from 26.9% to 34.9% between November 2025 and February 2026, while citations to profile pages fell from 33.9% to 14.5%.

A well-optimised static profile is no longer sufficient. What you publish is now part of your AI footprint. Your headline, About section, content, and consistency together form the signal AI reads.

One further Semrush finding is worth holding onto: 95% of the LinkedIn content cited by AI was original content rather than reshares.

LinkedIn’s own AI matters too

LinkedIn has been moving towards AI-led ranking and recommendation systems. One public research paper describes 360Brew, a 150B-parameter model built to handle ranking and recommendation tasks across LinkedIn.

Whether or not every feed decision runs through that exact model, the direction is clear: LinkedIn is moving away from simple engagement hacks and towards richer signals around relevance, identity, and expertise.

That makes coherence matter more. When your headline says one thing and your posts say another, the public signal gets weaker. When your profile, expertise, and posting topics reinforce each other over time, the pattern becomes easier to recognise.

That matters for AI visibility beyond LinkedIn too. The platform has been improving its structured headings, content architecture, and emphasis on time-stamped, expert-authored, conversational content. Those are the kinds of signals retrieval systems use when deciding what to cite.

5 things to fix now

1️⃣ Treat your headline as a keyword field

LinkedIn uses your headline as one of the primary signals that helps determine where your profile surfaces in search, including within LinkedIn’s own AI-powered search and in external AI tools. A headline that reads “CEO at [Company]” is searchable only for those two terms. A more specific headline that names your niche, your buyer’s problem, and the outcome you deliver creates more useful signals.

2️⃣ Align your About section with the topics you want to own

Your About section needs to reinforce the same expertise your posts are building. Write it for a search query, not as a biography. Use clear language about who you help, what problem you solve, and what changes as a result. Then check that the first 200 to 300 characters, the text visible before “see more”, answer the question: is this person for me?

3️⃣ Post original long-form content on your main topics

Posts in the 500 to 2,000 word range give you enough room to explain something useful, show how you think, and create original material AI tools can cite. Consistently sharing practical advice in mid-length posts gives both humans and machines more to work with.

4️⃣ Post consistently on the same two or three subjects

360Brew’s broader logic suggests topic consistency matters. Posting repeatedly on the same niche subjects helps signal expertise to both LinkedIn’s algorithm and AI search systems. Coherence matters more than volume.

5️⃣ Think beyond the profile page

LinkedIn’s AI-powered people search has expanded the role of natural language in how people are surfaced. A founder whose profile consistently frames them as someone who helps Series A fintech startups build investor relations can appear in search using exactly those terms, provided that signal is coherent and sustained over time. A list of credentials tells the algorithm very little about whether you are the right person for a specific query. A clear description of who you help and what changes as a result gives it something more useful to cite.

The question has changed

The shift from being found to being cited is becoming one of the most important visibility changes in AI search.

Even if LinkedIn still feels like a professional network you maintain, it is increasingly part of the professional knowledge base AI tools use when they decide how to describe and recommend someone.

If you build a consistent, credible, content-rich LinkedIn presence, you’re shaping the AI footprint that influences how you are discovered and described across the platforms your prospects use.

The question is no longer only: does my LinkedIn look good?

It is: what does AI say about me when someone asks?

👤 Melanie Goodman is a former Magic Circle lawyer turned LinkedIn strategist and founder of Trevisan. She helps professionals in law, finance, HR and other regulated industries build trust, attract clients, and get seen on LinkedIn without turning into an influencer cliché. She writes The Link Tank, a practical LinkedIn publication for consultants and service-based business owners. Connect with Melanie on LinkedIn.

The Link Tank™️

A warm welcome. The LinkedIn strategies I use to build authority, grow a newsletter & convert connections into paying clients. Every week, get practical, evidence-led LinkedIn tactics and know-how to credibly grow your LinkedIn presence.

By Melanie Goodman

By the end of the week, you’ve dealt with a lot and still feel like the real work never got touched. The thinking that moves the business forward got crowded out again.

, who writes All the Margins, has led through exactly that kind of pressure.

In this guest piece, she writes about what reactive leadership does to your judgement, your calendar, and the quality of the decisions your business gets from you. 👇🏻

Urgency takes the week before strategy touches it

You can lose a whole week to other people’s urgency and barely notice it happening. Someone is always awake, constantly working, and usually needs something right now.

By Friday, you realise most of your time went on dealing with whatever reached you first, not on the work that actually needed your judgement.

That matters more now because AI is making execution cheaper and judgement more important. If your week is filled by whatever feels urgent, you lose the part of the job that still needs you most.

I learned the hard way, leading large international teams, that without protected time every day brought its own urgent crisis and I lost the thread of what I was actually there to do.

Each one pulled me further from the work that actually needed me. A client fire drill on Monday. A team member falling behind on Tuesday. A data cleansing crisis. A third-party supplier failing again.

The cost hit in ways I didn’t see quickly enough. Two members of my leadership team were pulling their teams in opposite directions. I was so consumed by the daily fire drills that I spotted the problem too late, and the damage was already done.

By the end of some weeks I was running on empty. My decisions were reactive, shaped by someone else’s emergency rather than a direction I’d chosen.

On Monday mornings, when we sat down to set priorities for the week, I was carrying the residue of everyone else’s week into that meeting. The tail was wagging the dog.

The problem looked like a workload problem. It ran much deeper.

When you haven’t had space to step back, you stop thinking ahead and start managing whatever is right in front of you.

Reactive mode cuts off the kind of thinking that moves things forward. You deal with what’s in front of you and lose sight of where things are heading. That space isn’t going to appear on its own.

Once I realised the system is designed to consume your attention, everything changed. I became much stricter about where my time goes and less willing to give it away just because something feels urgent.

Research found that unfolding events consume around 36% of a CEO’s time. The real skill is knowing which events deserve your attention.

Here’s what that looks like in practice.

Protect the time first

No-meeting Fridays are a default for me, not an absolute rule. There are always exceptions.

Protecting Friday changes the whole week. It creates room for proper thinking and for conversations that are not squeezed between back-to-back calls. Walking helps too. Indoors or outdoors, it’s better for creative thinking than sitting.

This year I introduced office hours. It’s a recurring slot for the smaller asks that would otherwise arrive as mid-day interruptions: a ten-minute conversation, a yes or no, a sign-off on a logo design or a client response.

It takes around 23 minutes to regain full focus after an interruption. Putting smaller asks into a predictable slot gives the team a reliable rhythm and gives me back time to concentrate.

I also request agendas before meetings. Not to control the conversation, but to give myself time to reflect, think through my questions and the points I want to make.

Writing them down means I won’t get pulled into someone else’s agenda. My thinking stays clearer and I’m less likely to lose the point.

See where your focus is

A colour-coded diary sounds too simple to matter. It isn’t. When I look at my week in colour, I can see where my attention’s really going. That’s often not where I thought it was. A week that feels balanced rarely looks that way when you map it honestly.

My gym time goes in the diary too, even though it sits outside office hours. If it’s not scheduled, it won’t happen. I repeat that to my team regularly. The week will fill, and it’ll be the first thing to go.

If you don’t protect it, the week usually pushes it out.

Track your attention

Quarterly reviews, client deadlines, and team issues will always be more pressing than learning time. Without blocked time, the calendar fills before learning gets a look in.

I’ve found a way to protect it, though not yet where I want it. For now, it sits on Saturday mornings. It’s in the diary, it’s visible, and I’ve arranged other personal and family commitments around it to make it real.

That matters because it signals to everyone, including me, that this isn’t optional. I’m working to bring it into the working week, where it belongs. But it only gets done if you start where it fits.

The week won’t give you the space to think. You have to make it. Start with the calendar. Look at this week’s diary. Is it shaped by what matters, or by whatever got to you first?

👤 Daphne Mavroudi-Chocholi writes All the Margins, for people who are accomplished, stretched, and quietly asking themselves whether this is really as good as it gets. She has built and turned around businesses across neuroscience, AI, accessibility, and pet care, led teams, managed boards, and made difficult decisions without a manual. Connect with Daphne on LinkedIn.

All The Margins

All the Margins is for people with full lives who want more from them — without adding anything, overhauling anything, or pretending the answer is a morning routine. Twice a week on the small moves that change how your life feels from the inside.

By Daphne Mavroudi-Chocholi

Then a few months later, you realise it was doing more work than you thought.

Pressure makes bad cuts feel rational, and a bad cut rarely looks bad at the time you make it.

, the CFO who writes The Profit Lens, has seen that pattern up close more than once.

In this guest piece, she explains the difference between costs that protect the business and cuts that bleed later.

This is how to tell which is which before you find out the hard way. 👇🏻

There’s a mistake I’ve watched businesses make in every crisis I’ve lived through professionally. They cut without understanding what they’re actually cutting.

When pressure hits, most founders open the P&L, look for costs, and start cutting. It feels like action. In the short term, the numbers improve. EBIT recovers, cash extends, and the board is satisfied. Then, quietly, something else begins to break.

The cost that saves your business and the cost that damages it can look identical on a P&L. Both are line items with a number attached. Neither comes with a label that says “don’t touch this one.”

Understanding the difference matters before a crisis forces the issue.

The cut that looked rational

During a cost optimisation process I observed at close range, a leadership team made a decision that looked entirely rational on paper.

The business was under pressure because revenue was uncertain. The instinct was to reduce costs and protect margin, so the leadership team restructured the sales bonus scheme.

The new scheme was harder to achieve, rewarding growth rather than managed business. The logic was sound. What happened next wasn’t in the model.

The sales team, already navigating a difficult market, read the new scheme as a pay cut disguised as a restructure, and motivation collapsed.

The people who had the deepest client relationships, the ones who’d built the revenue base the business was trying to protect, started disengaging. Some left, some stayed but stopped trying.

In one quarter, revenue fell by half. It wasn’t the bonus scheme alone. It was what the bonus scheme communicated: that in a moment of pressure, the business chose its own margin over its people.

The signal travelled fast, and behaviour changed. Client relationships weakened and revenue followed. The cost of the old bonus scheme was visible. The cost of the new one didn’t show up until it was catastrophic.

The damage shows up later

This isn’t a professional services problem. I’ve seen the same dynamic in technology companies, retail, manufacturing, and early-stage startups. The business changes, but the mistake doesn’t.

A founder I worked with decided to cut a sales team that wasn’t delivering results. The logic was straightforward: three quarters of missed targets, no visible pipeline activity. The cost was significant and the return appeared to be zero.

What nobody measured was the conversion cycle in their industry. Six months after the team left, several of the prospects they’d been nurturing signed contracts with competitors.

The relationships had been built. The trust had been established. The work of eighteen months had been done. But the people who had done it were gone, and the business that benefited wasn’t the one that paid for the work.

The founder had cut a cost that looked like overhead. It was a driver. The damage didn’t appear on the P&L for two quarters. By then the connection between the decision and the consequence was lost.

The cut looked sensible that month. The revenue damage appeared later, when contracts that should have landed did not.

That’s how driver costs work. The saving shows up first. The damage arrives later, after the decision has already been made and explained away.

Not every cost is overhead

In twenty years of finance, I’ve come to think about costs in two categories that don’t appear anywhere in standard accounting.

Overhead costs support the business without directly generating revenue or client relationships. Tools, subscriptions, office resources, administrative functions, non-client-facing processes.

These costs matter. The business couldn’t function without them. But if you reduce or remove them, the revenue engine keeps running. Clients don’t notice. Relationships don’t break, and the pipeline doesn’t shrink.

Driver costs are directly connected to the business’s ability to generate and retain revenue: senior consultants who know client preferences, sales teams who hold relationship capital, marketing that fills the pipeline, and the specialist whose expertise is the reason a client renews.

These costs don’t just support the business. They are the business. When they’re cut, the damage shows up in revenue and retention, and the pipeline.

A driver cost is almost never obvious from the P&L line alone. A “consultant fee” could be overhead or the single most important relationship in your client base. A “sales incentive” could be a discretionary reward or the mechanism keeping your best revenue generators in the room.

The label tells you nothing. You have to ask the question.

Where cuts bleed

Across different industries and business models, certain categories appear again and again on the wrong side of the cut.

Sales support and business development activity

These feel discretionary when revenue is uncertain. They aren’t. Conferences, events, and relationship-building spend often show no visible return for months or years, and then a conversation from eighteen months ago becomes a signed contract. Cut this category and the pipeline doesn’t collapse immediately. It shrinks quietly over time.

Retention roles and people with institutional knowledge

The senior person who knows how every client thinks, what every internal process looks like, and where every relationship stands isn’t easily replaced. Their cost is visible, but their value usually isn’t. The moment they leave, the loss begins, but it often takes a year to fully appear.

Training and development spend

Cutting it feels safe because nothing breaks immediately. But a team that stops developing stops improving. In knowledge-based businesses, the quality of thinking is the product. When thinking goes stale, the work does too, and clients feel it.

Delivery quality investments

The tools, resources, and time that allow a team to do excellent work rather than adequate work. Cutting here cuts quality as well as cost. In markets where reputation drives referrals, quality is a driver cost.

Brand and trust-building activity

Content, thought leadership, community presence. These feel like nice-to-haves when cash is tight. But they’re often how future clients find the business in the first place. Cut them and the top of the funnel narrows. The effect appears six to twelve months later, when new business slows.

Ask these four questions first

When I’m working with a founder through a cost review, I ask four questions about every significant line item before I recommend anything.

Does this cost help us win work?

If the answer is yes, it’s almost certainly a driver. Anything connected to pipeline, business development, brand, or sales capacity belongs here.

Does this cost help us keep work?

Client retention in professional services is built on relationships, expertise, and consistent quality. Any cost connected to the people, tools, or processes that deliver these things is a driver.

Does this cost protect trust, speed, or quality?

Those three things are what reputation rests on in any service business. A cost that protects any of them is worth examining carefully before removing.

Would losing this cost show up later rather than now?

If the effect would appear in six months rather than six weeks, that’s a signal you’re looking at a driver, not overhead.

Start with the bank statement

The most reliable place to start a cost review isn’t the P&L. Open your bank statement and filter recurring payments, subscriptions, licences, and monthly commitments. List them.

You’ll almost always find costs you forgot you were paying for: tools adopted during a growth phase and never reviewed, licences bought for a team that has since changed, subscriptions that made sense eighteen months ago and no longer do.

These are almost always overhead. Start here. Reduce what’s not being used. Check licence volumes against actual usage. Ask vendors for discounts. In a difficult market, most will negotiate rather than lose the contract.

Then work through the remaining costs with the four questions above. Start with fixed costs, the ones that arrive regardless of revenue, then move to variable ones.

For each one, ask what would actually happen if it disappeared tomorrow, not what the model says. What would actually happen to clients, pipeline, or the people who deliver the work.

The answers will tell you what to cut and what to protect.

The spreadsheet won’t warn you

The founders who recover fastest from a difficult period usually know which costs can be cut and which ones will hurt the business if slashed.

Before you cut anything, ask what it’s actually doing for your business.

Overhead can usually be reduced without damaging the engine.

Driver costs can look similar on paper, but cutting them weakens the part of the business that brings work in and keeps clients.

The difference isn’t always obvious, but the cost of finding out the hard way is almost always higher than the saving that prompted the decision.

👤 Nadia Codreanu is a CFO with 20 years’ experience across corporate and founder-owned businesses. She writes at The Profit Lens about financial clarity, leadership under pressure, and the decisions that matter most for professional services founders. Connect with Nadia on LinkedIn.

The Profit Lens

By Nadia Codreanu

The best AI business may already exist

Wrapping AI around an existing workflow and charging a subscription is one opportunity. It’s visible, and in some cases it can work. It’s also often less durable than it looks. The model providers can add the feature themselves. Bigger products can fold it in. What looked like a business can get squeezed fast.

The more interesting opportunity is what happens when founders use AI to change the economics of work they already understand. That came through when I went back through 66 Millennial Masters interviews with founders already using AI in one way or another.

The money is often in keeping more revenue and making parts of the business viable that were too expensive to touch before. Smaller teams are part of that story, but not the whole thing. 👇🏻

Margin comes first

The clearest gains often show up in margin before they show up in fresh revenue.

A founder who already knows how to write, sell, analyse, research, or build can now cut a surprising amount of junior work out of the cost base.

The output still needs direction, judgement, and standards. The middle layer around it can get much thinner.

That’s one of the clearest ways founders are making money with AI right now: by keeping more of the revenue they were already capable of generating.

A business can hold the same top line and still get much stronger if delivery cost drops, outside suppliers matter less, and the same work gets done with less payroll around it.

That changes how much cash the business keeps, how flexibly it can price, and how long it can grow before it needs more people.

Anjeanette Carter described becoming “an army of copywriters” and deciding to stop working with writers, which cut a six-figure payroll.

That’s a story about protecting the revenue from a business she already understood by changing how the delivery gets done.

Hiring gets pushed back

Once margin improves, the next shift is headcount.

Some teams are still growing in the old way. A lot of others are holding the line on hiring because AI has changed what one capable operator can carry.

The business can do more before it needs another hire. Some roles do not disappear, but they arrive later.

If one strong operator can now carry work that once needed two or three people around them, the company can stay leaner for longer, delay hiring, and grow with a different fixed-cost base.

That changes when hiring makes sense and how much revenue the business needs before it can carry the next person.

Nick Holzherr said firms are no longer hiring marketing juniors in the same way. They’re keeping more senior people who can use AI well, which means they simply need fewer staff.

Carly Meyers believes huge companies will be built with teams of one, two, or three, with AI carrying a lot of the extra load those teams used to hire for.

Judgement still decides

This only works when the founder or operator already knows what good looks like.

AI can extend expertise. It does very little for people who did not know what good looked like before they touched it. Faster average work is still average work.

Anjeanette Carter was very clear that “if you don’t have expertise in a specific area, AI isn’t going to make it better. It’s only going to amplify your limited knowledge.”

That’s the point more founders need to sit with. AI can help when you already know what good looks like. It does far less when you don’t.

That matters commercially because it’s where the savings can turn false.

If founders mistake output speed for decision quality, they can ship more bad work, make more bad calls, and damage the business faster.

AI lowers the cost of production, but being wrong is still expensive.

Josh Payne said “human plus AI is currently still better than AI only in most cases,” and Ben Tasker described how an organisation replaced a call centre with AI, fired the staff, then hired them back a few weeks later because the system was awful.

The founders getting real value from AI are keeping human judgement on the parts that still need it. They’re using AI to absorb more of the repetitive work and the junior-heavy execution.

The human still decides whether the work is commercially safe and worth shipping.

Product is easier, winning isn’t

Once building gets cheaper, other parts of the business start to matter more.

Technical founders can now get further with less capital in less time. That opens useful opportunities. It also weakens the product as a moat when the product is the only thing the business has.

If more people can build something usable with AI tools, functionality alone becomes a thinner defence. Thibault Louis-Lucas asked where the value sits if you no longer need developers in the old way to create products. His answer is distribution and storytelling.

Simon Jenner explained that when build costs fall hard enough, niches that were once too small to matter start becoming viable. A business that made no sense at one cost structure can look very healthy at another.

When building gets cheaper, more founders can get to a working product. What becomes scarce is the ability to earn attention, build trust, and turn a functional product into something people actually choose.

The moat moves away from pure build difficulty and towards distribution.

The graveyard has money in it

Some of the strongest AI opportunities are sitting inside neglected assets that humans stopped touching because the economics were too weak before.

AI lowers the cost of revisiting things that were already there but no longer worth a person’s time.

Ryan Carruthers’ example was sales. He said AI is the best place to start there because you can test it on dead leads. If the process fails, the leads were dead anyway.

In his case, those systems had hundreds of conversations, booked sales calls, and did it in a place where the brand risk was low because the asset had already been abandoned.

It’s an operator looking at wasted value and realising the cost of revisiting it has changed. Yehong Zhu showed the same commercial instinct in a different market through licensing. She’s building a business around the way AI changes the value of media rights and data access.

The tool creates a new buyer. The founder changes what can be sold.

The premium moves elsewhere

The smartest founders are using AI to change the economics of work they already understand. They’re cutting waste from familiar work and reopening parts of the business that were too expensive to touch before.

If a product gets cheaper to build, the premium shifts toward trust, distribution, and judgement. Trust matters more when polished slop is cheap. Distribution matters more when a product is easier to copy.

Judgement still decides what is worth shipping, and domain understanding still decides what anyone will pay for.

Gary Das pointed to collapsing trust. Freddie Pullen pointed to the value of long-form human formats that still feel unmistakably real. Those are clues about where price and loyalty can still hold when AI has made the underlying production cheaper.

Founders who already know how to create value can keep more of it, do more with fewer people, and make businesses work that would not have made sense under the old cost structure.

The lasting opportunity sits inside work that already has a business underneath it. That’s where more of the value stays.

Related conversations:

The harder part is what people learn from you in the smaller moments. It’s what I learned on my own the hard way.

A bad reaction after one difficult update can teach the team more than any leadership principle you talk about.

, co-founder of Kestryl Edge, works in that space between behaviour, pressure, and team dynamics.

His guest piece gets into what emotional intelligence looks like when it leaves the leadership vocabulary and shows up in how people actually behave around you. 👇🏻

Your behaviour trains the team

Low-trust leadership usually shows up fast.

It creates dependency and makes people think twice before raising problems. Very quickly, the team learns that self-protection matters more than honesty.

People stop speaking plainly and start managing the boss instead of the work. They protect themselves before they protect standards, and problems get softened before they reach the surface.

Trust is set from the top

Trust matters because it shapes whether people feel safe enough to do the work properly.

Low trust doesn’t stay contained in one difficult interaction. It changes the whole environment around a team. People become hyper-vigilant. They start reading tone into emails and second-guessing how to say things safely.

Soon the day fills with small exhausting questions: How much do I say? Is this safe to admit? Will this be used against me? Over time, that wears people down. Judgement gets suppressed, creativity narrows, and the body carries the cost as well as the work.

A bad boss can have more influence over your health than people want to admit. That sounds extreme until you’ve worked for someone who keeps you in a permanent state of tension.

The body doesn’t always know the difference between a real threat and a workplace one. It just responds.

A team can’t do its best work while it’s constantly scanning for danger.

Power matters, and teams adapt around the person who sets the tone. That’s what makes emotional intelligence practical in everyday work.

Where trust is built

Leaders who know themselves understand how their mood changes a room.

They can feel the moment when they’re reacting rather than responding, which makes them less likely to turn defensiveness into policy or tension into tone.

People feel the difference between a leader who knows themselves and one who doesn’t, even when they can’t name it.

That’s where self-regulation starts to matter. Can you receive hard news without punishing the person who brought it? Can you stay steady in a tense meeting? Can you pause before turning irritation into damage?

Teams trust leaders they can anticipate. If your reactions are volatile, people start editing themselves around you. If your reactions are measured, they start bringing you the truth faster.

This is one reason I write while people talk. Notes slow me down and keep me listening a little longer than my instincts want to. It is a small thing, but trust is often built in small things.

When leaders regulate themselves well, teams learn that bad news can be raised without explosion and that mistakes will be examined rather than used for blame. Those moments do not just make people feel better, they make the work better.

Empathy matters for the same reason. It’s a form of attention, the discipline of not rushing past someone’s actual experience because you’re more comfortable with your own interpretation of it.

People trust leaders who try to understand their reality before judging it, who notice what isn’t being said, and who respond to the person in front of them rather than just the role they happen to occupy.

But empathy on its own is still not enough.

Behaviour teaches the team

That’s what emotional intelligence looks like in practice.

It isn’t jargon, posture, “clarity”, “alignment”, or any of the other polished words leadership culture likes to hide behind.

It means knowing yourself well enough not to let your own volatility run the room and regulating your response so other people don’t have to manage it for you.

It also means listening closely enough to understand what is actually happening for someone else and following through in ways that prove your behaviour can be trusted.

Low-trust leaders create low-trust teams because people learn from behaviour, not declared values.

When your behaviour teaches people to stay quiet, protect themselves, and hide the truth, that is what the team becomes.

When it teaches them that they can speak honestly, think clearly, disagree safely, and recover from mistakes without fear, that becomes the team too.

That’s why emotional intelligence matters: leaders decide, day after day, whether people around them feel safe enough to think and work properly.

What looks like high standards and hard work can easily become the reason your business stops growing.

When you build from scratch, being in the middle of everything feels normal. It’s part of why the business exists.

Then the company grows, the work changes, and you stay just as central because it still feels like the safe, responsible thing to do.

You might think the answer is more people or better systems. Often, it’s how much still runs through you.

After a while, that stops being a habit and starts becoming identity. The more Millennial Masters interviews I’ve done, the more I’ve recognised parts of this in myself.

Gary Das realised he was the problem.

“I was the bottleneck. The business had almost outgrown me. I hadn’t grown with the business as a person, in my mindset, my behaviours, and my habits.”

Your company can outgrow your old way of operating before it outgrows your ambition.👇🏻

You stay too central

This usually builds slowly, as you stay close to everything. Harry Sanders said it’s a symptom of founders who want to “know everything, touch everything, feel everything.”

Businesses like that stay around the same size because the founder never stops acting like the centre of gravity. Every decision routes back to you. The team waits for your reaction before moving.

You drop new ideas into work that’s already moving because being involved still feels like your job. That’s why the habit lasts longer than it should.

When you hold on to the version of yourself the early business needed, the strain shows up before the systems problem is even obvious.

Standards are just a cover story

This is where you can hide. You tell yourself the issue is quality or standards.

Gary’s attitude was that “no one can do it as good as me,” which made it easier to step back in instead of training people properly.

That can sound sensible when you’re the one doing it. But the harder part is letting someone else not do it your way.

Damon Flowers said founders often want staff to do things exactly the way they would do them, and they struggle to accept handing over a task and getting 60% of the result back while someone learns.

Founders often can’t tolerate that stage, so they step back in because fixing the problem still feels like their job. All that does is teach the team to hand problems back to you.

You teach the wrong habit

Damon’s point is what your behaviour teaches the team. If you take the work back every time it comes back imperfect, you train people in exactly the wrong way.

They learn they don’t need to finish the hard part because they know it can always come back to you. After a while, the whole company starts working like that.

Then you complain that nobody can take anything off your plate, while teaching the business that it still cannot function without you.

You call it high standards, but the team experiences a system where ownership never settles anywhere because you keep pulling it back in.

That’s one of the easiest ways to build a company full of capable people who never become fully capable inside your business.

The role changed, you didn’t

Tyler Dunagin uses an American football analogy: you start as the all-star player, move into coaching, and end up in the owner’s box.

That’s a tough change when your identity was built around being the only one who could move things forward through effort.

Harry Sanders said realising he wasn’t a good manager was an ego hit. “You can either be rich, or you can be king. You can’t be both. The king is the person that everyone loves, that comes in, says hi. Being rich is building the organisation that doesn’t need you.”

Delegation is only part of it because your role changed. The business needed you to build the structure around the work and let other people carry more of it.

Part of you still wanted the emotional reward of being the all-star player. That’s normal. It also keeps the company smaller than it should be.

Then you resent the business

Joy Zarine’s example is painfully recognisable. You’re on holiday, laptop by the pool, answering emails, annoyed, thinking this isn’t why you built a business.

You keep yourself central because it feels safer, then you resent the business for still needing you. You complain that you can’t switch off while still holding onto the habits that make switching off impossible.

Joy’s point was that this is on you. The frustration is real, but the structure underneath it is self-created. You need people around you, and you have to let them mess up, as painful as that is.

Emma Mills was told more directly that she was stuck in her ways. She runs a personal assistants business, but didn’t have a PA of her own.

Her mastermind facilitator pointed out she was coaching other owners on leverage while still being the bottleneck in her own business.

You can usually see the pattern in other people long before you see it in yourself.

When pushing harder stops working

Mike Jones lived the darkest version of this in his CrossFit business. He tried to push through scaling, but the way he was running it started taking a real toll on his life.

He worked more and more, the business got on top of him, and in the end his brain couldn’t cope. The pressure built until he shut down a good business.

That strips away the idea that this is just a productivity problem. You can carry the wrong structure for a while and still keep the business alive.

There’s still a cost, and it can be big enough to take the whole thing down.

What changes when you step back properly

James Fleming admitted he “was getting frustrated, and the excuse I was making was… it’s your fault,” pointing fingers at the team.

He realised “I had to change” when one of his senior leaders called him out for pushing too hard.

He started stepping back and saying, “I don’t know, but I know you guys do,” passing ownership back to the people he’d hired.

That matters because the role changed. Your job is no longer to fix everything yourself. It’s to set direction, hire well, and trust people to do the work to the standard you expect.

Without this identity shift, the business still runs through you, even after you’ve built the team.

If you keep stepping back in every time something is imperfect, you stay the bottleneck. You’re teaching the company that it still can’t move without you.

If that never changes, the business stays stuck with you at your limit.

Related conversations:

That sounds like an advantage until you realise it also makes it easier to waste time and money on the wrong idea.

, partner at MGV Capital and founder of Alpha Bytes, where he writes about startups, investing, technology, and building companies, has seen that from both the operator and the investor side.

In this guest piece, he writes about why founders need better filters now that building is cheap and exploration matters more than ever. 👇🏻

The real skill is choosing what to build

We’re idea machines by nature. That’s never really been an issue.

The problem is that most ideas, even the good ones, aren’t always worth building.

They could be worth exploring. That distinction matters more now than it used to.

The filter is gone

In the old world, scarcity did the filtering for you.

If an idea needed six months and serious capital to prototype, you stress-tested it before writing the first line of code.

You asked hard questions early because the cost of being wrong was enormous. Time, capital, and effort were brutal filters, but honest ones.

That world is disappearing quickly. AI has collapsed the cost of execution in ways that would’ve sounded absurd even a few years ago.

You can spin up an MVP in a weekend and get something usable in front of people almost immediately. The friction that used to kill bad ideas early has gone.

That sounds like progress, and in many ways it is. But it also makes it easier to waste time building the wrong thing.

The bottleneck has changed. It’s no longer “can we build this?” It’s “should we build this, and should we build it now?”

Exploring is not building

After being on both sides of the fence as an entrepreneur and an investor, I’ve noticed there’s a stage between having an idea and shipping it.

That stage is exploration. It’s where you test an idea before putting real resources behind it. No code. No hires. No capital. Just the cheapest due diligence you can do.

The problem is that exploration is messy by nature. Ideas sit in your head, in random notes, in voice memos that make no sense the next morning.

That’s where a system helps. Instead of relying on gut instinct, it gives you a repeatable way to capture, evaluate, and either park or promote ideas.

The goal is to sequence ideas and finding the ones that are actually right for you.

Some ideas are good on paper and still wrong for the person trying to build them because they don’t match their experience, skills, or ambition.

A simple way to do it

I’ve been running a version of this for a while, and it’s changed how I think about new projects. Here’s the framework stripped back to the essentials:

1️⃣ Capture everything. Write every idea down. No filtering. A note, a voice memo, whatever works. The point is to get it out of your head. Unsorted ideas create cognitive load. A log helps with that.

2️⃣ Let it sit. I usually do 72 hours. Some people do 48 hours, some a week. What matters is that you don’t act on it within 24 hours. Excitement is a bad filter. The ideas that still feel interesting after the dopamine fades are the ones worth exploring.

3️⃣ Score, but do not judge. For the ideas that survive, run them through four questions. Does this solve a real problem I can explain in one sentence? Do I have a unique angle or unfair advantage? Is there any sign of demand, even anecdotal? Could I explain this to a stranger in 60 seconds? Anything below your threshold gets parked.

4️⃣ Explore on paper. For ideas that pass the filter, do a focused two- to four-hour research sprint before you build anything. Map the competitive landscape, look at existing solutions, and find communities already talking about the problem. AI can help you synthesise information, but it shouldn’t make the decision for you.

5️⃣ Define the kill criteria. Before you go any further, write down the specific conditions under which you’d stop pursuing the idea. If you can’t articulate what would make you walk away, you’re probably not being honest about the opportunity.

6️⃣ Keep a live repository. The ideas that survive aren’t a to-do list. Revisit them monthly. Look for recurring themes, overlapping audiences, and complementary concepts. Sometimes the best product isn’t a single idea in the log, but the overlap between several of them.

7️⃣ Promote one. Park the rest. When you’re ready to commit real resources, pick one idea and move it into formal validation. Build a landing page, set up interviews, create a waitlist, or talk about it publicly. The rest stay parked.

The trap you need to see coming

This system has a failure mode, and it’s a seductive one. Too much exploration without commitment is just procrastination.

The system is there to help you decide. If you’ve been exploring the same idea for six months without moving it into validation, that’s probably your answer.

There’s also the temptation to get attached to having a rich idea log without ever acting on it.

The log is there to help you move. To stop it turning into a trophy case, set a cadence, quarterly at minimum, where you force a promote-or-park decision on your top ideas.

AI can make this trap worse. Because it can generate research, competitive analysis, and prototype mockups on demand, it’s easy to confuse activity with progress. A lot of output can still mean very little progress.

The real advantage now

Humans will keep generating ideas. That part isn’t changing.

What’s changed is the cost of acting on them, and with it, the cost of acting on the wrong ones.

In the AI economy, speed alone won’t be enough. The difference is knowing what deserves real commitment and what should stay in exploration.

👤 Barry Winata is a partner at MGV Capital and writes Alpha Bytes. Before moving into investing, he spent nearly a decade as an operator across engineering, growth, recruiting, product, operations, and sales, helping startups grow into successful companies, including businesses acquired by Verizon and others backed by investors such as SoftBank, Intel, and Samsung. Connect with Barry on LinkedIn.

Alpha Bytes

The pragmatist's anti-hype guide on AI and its impact on you, your career, business, culture & society.

By Barry Winata

I’ve seen it and I’ve done it, usually when you’re stretched and trying to keep too many things moving at once.

Marketing feels vague and product feels too technical, so you bring in outside help to “handle it”.

That can seem sensible, like you’re finally out of the weeds. Sometimes it is.

A lot of the time, though, you’re using outside help to avoid learning a function you still need to understand.

That’s where the trouble starts. You hand away part of the business before you understand it well enough to judge it or spot when it’s going wrong. 👇🏻

You skipped the hard bit

Melissa Kwan ran into this twice. She kept hiring agencies to do the hard work she didn’t want to do, then got poor results because she didn’t even know the basics of marketing.

She never really took that on because she told herself it wasn’t for her. In the end, she put “the most important revenue role” in someone else’s hands.

If the function is still a black box to you, you don’t really know what you’re buying.

You don’t know how to write the brief, how to interview for the role, or whether the person in front of you is excellent, average, or just saying the right words.

Melissa realised that “you cannot hire someone to do a good job if you don’t even know how to interview for them… It’s not that they’re not good. It’s that I don’t know how to direct them.”

You think you’re fixing a capability gap. What you often create is a visibility gap.

A product doesn’t stay finished

Melissa ran into the same problem again when it came to product. The issue was never just agencies. It was handing away a core function.

She hired a dev shop that was meant to cover everything: PM, developers, technical lead, the lot. The deal was milestone-based. It sounded sensible, but the product didn’t really work, the shop kept burning time, and the moment real users arrived the gaps started opening everywhere.

Once a product is live, it keeps breaking, changing, and needs fast decisions. A dev shop isn’t built for that kind of urgency. “A dev shop will never be able to fix that. They’ll be like, yeah, maybe next week.”

A project can be outsourced. A live product usually cannot be treated like one.

If what you’re handing away needs constant judgement, fast iteration, and real ownership, you cannot afford to have all of that understanding sitting outside the company.

Agencies optimise for the incentive

Gary Das found a similar issue in his marketing spend. He was paying marketing companies to get leads for his mortgage business, and all they cared about was getting leads because that was the metric they were paid on.

They weren’t bothered if anyone answered the phone, if those leads became appointments, or if they ever turned into clients.

That isn’t a moral failure. It’s an incentive structure. Outside help usually optimises for the metric it gets paid on, and that metric can sit one step away from what actually matters.

If you don’t understand the function well enough to see that gap, you can burn a lot of time and money while everyone involved still claims the work is performing.

You can outsource execution, but you still have to own the result.

You still need a clear view

Damon Flowers gets to the practical fix. If you’re hiring for something technical and you don’t have the background to judge it, bring in someone independent who does.

Tap your network and get somebody to shadow the work, review it, or tell you whether what’s being built is what should be built.

You don’t need to become the world expert in every function. You just need enough understanding to know what good looks like. If you don’t have that yet, you need someone close enough to the business who does.

The damage starts when you lose visibility. When something gets blindly handed off, you start losing sight of the pipeline, the cost, and the key metrics that tell you whether the engine is healthy or leaking.

The problem is outside help plus founder blindness. Melissa found the fix almost by accident. The outsourced dev team kept burning through time and money without getting the product working, so David, who had been her boyfriend for seven years, stepped in pro bono to help her get unstuck.

Once she saw what he could do, it made no sense to keep paying an external agency for poor results. She brought him in as CTO and co-founder, sorted the equity, and now he runs the remote engineering team in Vietnam.

That also gave Melissa room to focus on what she needed to own herself. She could bring marketing in-house, rework the messaging and positioning, and stop outsourcing a critical part of the business she had already spent years avoiding.

Some knowledge needs to stay inside

“Do you want to have your own army, or do you want to pay mercenaries?”

Yannik Schrade’s point is that you need to think hard about what your moat actually is. If the team’s skill set and knowledge of how to solve the problem are part of that moat, building that expertise internally matters.

That’s why he’s in favour of having your own team, where the capability compounds inside the business instead of leaking out through outsourced relationships.

A lot of founders miss this because they treat everything as a resourcing problem. Some tasks are just execution. Others are where your edge starts to form.

If the function shapes revenue, product quality, or how the business learns, handing it away too early can do real damage. It can stop the company from building its own judgement.

Where outsourcing actually works

Liam White is a good example of where outsourcing does make sense. He realised that if his sauces business wanted to stay lean, it made more sense to find a manufacturer and keep the internal focus on sales, marketing, and product. The manufacturer had expertise it would’ve taken years to build in house, and outsourcing let them scale faster.

Harvey Armstrong and Sam Holmes are beer lovers, not beer makers. They work with professionals on producing Prime Time, keep control of the brand, and use outside expertise to scale without pretending they need to own every technical detail from day one.

Joe Averill makes the same point with fractional talent. A startup may not have the money for a full-time CFO. Fractional support gave him access to higher-level capability without the full-time cost.

So this is not an anti-outsourcing argument. It works when you are buying specialist help or capacity you do not need to keep inside the business. It breaks down when you use it to avoid understanding a core function.

The rule that matters

If the work is core to revenue, product quality, customer understanding, or your future moat, you need enough understanding to judge it.

That doesn’t mean doing everything yourself forever. It means not handing away the part of the business you still don’t know how to see clearly.

The expensive mistake isn’t paying too much. It’s handing away a function before you know how to see it clearly.

That’s why AI feels so powerful to solo founders, consultants, and service businesses right now.

The catch is that when everyone has access to it, the advantage shrinks fast.

of Unpromptable is back with a guest piece on what actually keeps a business distinctive once the tools are no longer the story. 👇🏻

The new AI moat is harder to fake

A year ago, saying “we use AI in our workflow” turned heads. Today, it feels like every service business is running some flavour of ChatGPT, Claude, or a no-code workflow bolted together over a weekend.

The differentiator disappeared. “We use AI” now carries the same weight as “we use email.” It’s the baseline. And if your competitive advantage was speed or cost, AI just handed that same advantage to everyone else in your market.

The uncomfortable truth for most digital business owners is that the only durable advantage left is becoming the kind of business that AI can’t replicate.

I call this unpromptability. It’s the part most people miss when they talk about AI advantage.

Blending in gets expensive

Most businesses haven’t felt the full weight of this shift yet. They’re still riding the early efficiency gains, with faster turnarounds and lower overhead.

The gains are real, but they’re also temporary, because every competitor has access to the same tools. Speed used to matter when it was rare. Once it becomes baseline, it stops carrying much pricing power.

You can’t differentiate. When everyone’s output starts looking and sounding alike, your prospects can’t tell you apart from the next option in their inbox. You become interchangeable. And interchangeable businesses don’t command attention, loyalty, or premium pricing.

You race to the bottom. Without meaningful business differentiation, the only lever left is price. You start undercutting. Your competitors undercut back. Margins thin. The work gets commoditised even if the service behind it is good. AI didn’t cause this dynamic, but it accelerated it by a decade.

Your clients hold all the leverage. When everything looks the same, buyers default to convenience. They shop around faster, switch providers more easily, and negotiate harder. The relationship you thought was solid becomes transactional overnight. They’re loyal to the outcome, and if someone else can deliver it cheaper or faster, they will.

This is the AI commodification trap. The way out is becoming harder to replace.

That usually starts with diagnosing the real bottleneck properly. A lot of businesses think the problem is speed, so they use AI to push out more work. Then they end up with faster average output and the same strategic weakness underneath it.

The advantage now is knowing where the tools actually help, and where they just make the sameness worse.

What competitors can’t copy

What makes a business harder to replace is not the tool stack. It’s the part competitors can’t easily copy: a clear point of view, real relationships, useful assets, and proof that the work delivers.

Marketing matters differently once everyone has access to the same tools. It stops being about promotion and starts becoming proof. The businesses that hold up are the ones whose work carries enough specificity that it cannot be mistaken for the next polished option in the market.

The clearer your thinking and the more visible your body of work, the harder it is to become interchangeable on price alone.

Relationships matter because AI can imitate your tone, but it can’t build rapport, trust, or the kind of connection that turns one client into a long-term buyer or referrer.

Proof matters more too. In a market full of polished claims, case studies, repeat clients, and visible outcomes do more to separate a business than another clever line about AI ever will.

That is the real skill here. AI made building easy. Building something people trust, remember, and come back to is still the harder part.

👤 James Presbitero Jr. is the founder of Unpromptable, where he helps founders and service businesses build AI systems that make them impossible to replace. If you want to know where your business stands on the irreplaceability spectrum, take the free AI Readiness Diagnostic. It takes 3 minutes and gives you a clear picture of where to start.

Unpromptable.

Helping founders, creatives, and professionals become irreplaceable in the Age of AI || Subscribe: forever free 2x deep dives, community events, and free Active Prompt Vault (+ more AI & marketing tools).

By James Presbitero

There’s a point where underpaying yourself stops being discipline and starts affecting the person making the company’s most important decisions. That’s when the founder becomes the weakest part of the business.

I heard it put in different ways by Millennial Masters guests, and most of them treated it as part of the deal. I did that for too many years, until I realised what it was costing.

Melissa Kwan also used to think founders should get paid last and least. After years of doing exactly that, she changed her mind. “The biggest variable to success is the founder,” she told me, and “if you are constantly in survival mode… you cannot make good decisions for your company.” The real issue is what that does to the founder.

Before eWebinar, she described almost ten years of “pure suffering”, paying herself barely enough to get by well into her 30s, living with multiple roommates, picking the cheapest thing on the menu, or meeting friends after they’d already eaten because she couldn’t afford dinner with them. That’s what bootstrapping often looks like in real life.

It’s easy to package that kind of sacrifice as part of the job. It sounds different when you look at what people are actually giving up.

James Fleming went from a highly paid executive job to selling his watches while his wife sold off her gold just to keep the marketing going. He had days when he didn’t know how he was going to put fuel in the car to get to meetings. Emma Mills ended up remortgaging her house to sort out cash flow after getting too optimistic about what the business could support.

Tyler Dunagin pulled money out of his savings and retirement account and paid himself a pittance for the year. Andrew Steele pointed to another version of the same problem: founder pay that’s too low or too erratic can leave you unable to get a mortgage at all.

People blur the line between lean and self-destructive when they talk about “lean”. The word sounds smart and controlled. The reality is often much uglier.

Yota Trom says scarcity leaks. If you’re operating from that mindset, people feel it, even if you never say it outright. Her question stings: who wants to be part of a culture like that? She pushed it further by asking how you can be generous with other people if you’re not generous with yourself. That moves the conversation out of personal finance and into company culture.

The overlooked bit is what chronic money stress does while you’re still trying to lead well. It can show up as caution that looks sensible but is really fear. It can make you more defensive, tighter with people, more hesitant to hire, slower to spend where it matters, more transactional, and less trusting.

It also pulls you into short-term thinking. When money stress is constant, you start solving for immediate relief instead of better long-term decisions. You might call it prudence, but the team experiences it as pressure. And the people you most want around you usually don’t stay long in a culture shaped by private financial panic.

Damon Flowers pointed to another part people tend to leave out. Founder money stress is lonely. You can’t really take it to staff because then they start wondering whether they should look for another job. You can’t fully take it home either. So you end up carrying it in your own head, and it gets heavier. That isolation matters because people don’t make their clearest decisions when the financial pressure is constant and private.

Peter Watson made the case for taking less and keeping more inside the business. Founders should work out the minimum they can live on, then take that, and leave the rest in the company. That can give a business breathing room. It can leave you with more room to survive, invest, and move when opportunities appear. Some founders do win that way.

But Peter also shows the trade-off. Even later, after the business had improved, he was still pricing purchases against an old £5-an-hour wage. He describes how hard it is to “unlock” that after years of living so frugally. Extreme discipline can help a founder survive. It can also rewire their relationship with money in ways that don’t disappear once survival is no longer the issue.

A founder taking less for a while can be rational. A founder who stays so underpaid that they become distracted, anxious, reactive, isolated, resentful, or mentally worn down is no longer helping the company.

That was Melissa’s point. If you’re constantly in survival mode, you cannot make good decisions for your company. Once the strain starts affecting judgement, it’s already affecting the business.

None of this means founders need luxury first, or that every business can pay market-rate salaries from the start. It also doesn’t mean frugality is fake or that every rough period is a mistake.

It means the founder should not be the part of the business quietly breaking while everyone pretends it’s discipline.

The founder doesn’t need comfort first. They need enough stability to think clearly, lead properly, and keep building. Without that, the business pays for it.

Related Millennial Masters conversations

You push hard, things move, users show up, the dashboard looks alive, and it all feels like proof you’re getting somewhere.

That’s where founders start lying to themselves. Not because the numbers are fake, but because they are reading the wrong ones.

You can end up building confidence off the wrong signals and run the business into the ground without realising it until much later.

That’s why I invited to write this. He’s the founder of Inkjin, Europe’s tattoo marketplace, and has built startups across adtech and crypto.

What he gets into here is one of the easiest traps founders fall into early: mistaking movement for traction. 👇🏻

You have movement, not traction

The dashboard can look healthy and still tell you the wrong story. I’ve had to learn that for myself more than once.

Signups keep going up. Engagement is “solid.” Retention is just good enough to keep you optimistic. So you proceed: extend the roadmap, hire the next team member, raise the next round of funding. Three months later, growth has stalled, your burn is up, and you still feel no more certain.

The data wasn’t broken. The problem was what you thought it meant.

Motion can look like traction

Most startup dashboards are built around activity: who came, how long they stayed, how often they came back. A user can look, like, and come back, but still not be committed to you. This isn’t traction, it’s a warmup.

This gets worse when a startup has early momentum. Your first users came through word of mouth, a well-placed tweet, or your own relentless and very personal efforts to drag potential customers through every barrier the product had to offer.

That momentum is real, and it matters. But it’s also borrowed: from your marketing, your distribution, and your own effort as a founder. It only lasts as long as you keep pushing it.

The dashboard can be distracting because it never asks the harder question: Does anyone care enough about this product to make a serious investment in it?

The signal you actually want

Users will write to you to complain that the product is broken. They’ll ask for features that remove friction from their daily lives.

When something doesn’t work, they work around it instead of leaving. They email after you go quiet. They keep using it while they wait for a response.

The better metric is inconvenience. Would users be truly inconvenienced if you disappeared from their lives today? Not annoyed. Inconvenienced.

We misread our own product

Inkjin, our startup, helped users discover and connect directly with tattoo artists. We had very solid retention numbers. Users were spending a lot of time in the app, coming back repeatedly, looking through tattoo designs and artist bios.

They were doing what they needed to do, but the real decision-making was happening beyond the app: contacting artists, discussing designs, making final agreements.

Inkjin was helping people get there, but it hadn’t become the owner of commitment. The dashboard measured consideration at scale. It could not measure where the decision was made.

The harder signal came from artist feedback. When we asked artists who had contacted them and who had actually gone on to get tattooed, the answer was clear: we estimated that 20 to 30 monthly bookings were coming from Inkjin discovery and being completed outside Inkjin. The decisions were being made. We just weren’t tracking them.

So we changed our stance. We stripped down the dashboard. We stopped asking how many people were using it. We started asking who made a significant decision, who zeroed in on an artist and committed to an action, and who came back to that same decision two or three times. We changed how we tracked what counted.

The real traction test

Look at the past 30 days and, without hand-waving, answer three questions:

• Who actually acted on purchase intent?

• Who returned to the same decision a few times?

• Who could not do it without you?

If you can’t answer those clearly, that’s the signal. You aren’t measuring traction. You’re measuring motion. I run this test on our own product, and it makes me uncomfortable every single time.

If a company you’re evaluating cites daily active users, time on site, signups, a growing waitlist, or media mentions, you’re looking at motion.

Ask instead: who is dependent on this? Who gets upset when it breaks? What do users do tomorrow if this disappears?

If the answers are vague, the traction is borrowed. If the answers are specific, with named users or a concrete alternative, you’re getting closer to something real.

What this mistake costs

Founders who mistake motion for traction start making expensive decisions too early.

They add headcount, stretch the roadmap, and spend money as if demand is already there. It isn’t. The product still has attention, activity, and maybe even growth. What it doesn’t have yet is enough commitment to support those decisions.

Then the story turns defensive. “We had users but they didn’t convert. We just need more time. It’s a little early.” By then, the spend is already in the business and the runway is running out.

That’s how you end up building for demand that was never really there.

👤 Ilias Hatzis is the founder and CEO of Inkjin, Europe’s tattoo marketplace, and writes The Leap of Faith, where he shares what building startups has taught him about traction, growth, and the mistakes that compound. Connect with Ilias on LinkedIn.

The Leap of Faith

Thoughts and things I've learned building tech startups.

By Ilias Hatzis | Entrepreneur

Luck matters, yes. It just gets blamed for too much while the ordinary reasons companies fail are sitting there in plain sight.

That’s the argument Bernie Bulkin makes in his new book Why Start-Ups Fail, which I’ve been reading this month.

When you’re in the thick of it, with tech, staff, product, and constant questions all going wrong at once, you do not always see it clearly. Some distance changes that. 👇🏻

Failure gets dressed up

Failure gets talked about as courage, bad timing, or the cost of trying something ambitious. Founders get to sound brave and investors get to shrug off dead companies as the cost of backing the future.

Bulkin is critical of the venture model here. A fund does not need a portfolio full of strong, well-run businesses. It needs one or two huge outcomes big enough to pay for a long list of write-offs.

That leaves founders in a different game from the people backing them. You are trying to build something that survives. A fund built around home runs can ignore a perfectly good business if the upside is not big enough. If it goes wrong, you are the one left living with the consequences.

He also shows how easily this gets cleaned up after the fact. Bad decisions start sounding like ambition. Failure starts sounding like proof you were serious, when often it was just bad judgement with better PR around it.

Where the boring damage starts

One of the easiest mistakes to miss is the gap between proving something can work and building a business around it.

Something works in a lab, a prototype gets built, a demo lands, and people act like the hard part is over. It usually is not. Science proves something can work. Engineering turns it into something repeatable, reliable, affordable, and scalable.

Building one impressive thing is not the same as building a system that can produce ten thousand of them properly.

The same thing happens in the market. Founders love a big TAM, but the harder question is how the market actually works. A giant market means very little if you do not know who actually buys, how they buy, who sits in the middle, and what it takes to get them ready to buy.

Enterprise sales are no different. Founders love talking as if they are going to sell straight into Shell, Ford, or another giant incumbent. In practice, a lot of those industries buy through procurement chains, tiered suppliers, and middlemen. The buyer is often a lot less glamorous than the logo on the pitch deck.

Then they burn months in meetings and pilots while nobody buys. A big market does not help much if your company cannot actually reach the buyer.

The money usually goes wrong earlier

After 21 years in venture capital, Bulkin knows running out of money is usually not the whole story.

By the time the money runs out, the real damage has usually been building for a while. The company was underfunded, or the founder kept cutting the wrong things and calling it discipline.

A lot of founders tell themselves they are being disciplined when they raise too little to avoid dilution. Then they spend the next year stuck in fundraising mode, preserving a larger share of a weaker company while starving the thing they were supposed to be building.

Work that looks non-essential gets pushed back until you realise it was holding the structure up. IP protection gets delayed. Market work gets skipped. Nobody gets a proper grip on the numbers early enough. The company starts penny-pinching and ends up broken underneath.

He makes the same point with cash timing. Founders model salaries and rent, then act surprised when inventory timing wrecks them. Paying suppliers long before customers pay you is part of how the business works. If you do not understand that gap early, growth can make the company weaker, not stronger.

Valuations create another version of the same problem. High paper valuations can narrow the ways out, inflate expectations, and turn sensible outcomes into disappointments for the wrong people on the cap table. That kind of valuation can make the business much harder to save later.

Then the founder becomes the problem

This is the bit that can sting. The founder who is right in year one is often wrong by year three. What gets a company off the ground often starts causing problems later.

Bulkin’s founder types are easy to recognise. The technical founder who never really becomes someone who can run the business. The fundraiser who is stronger in pitch rooms than inside the company. The arrogant one who cannot hear challenge.

Once the company gets bigger, those weaknesses stop being quirks and start doing damage. Startups still give too much credit to confidence, visibility, and a good story long after the business needs something else.

It’s easier to keep telling a story about a bold founder changing the world than to admit the company needed someone different to lead it and never got that.

Most startups die because the boring bits were not taken seriously enough. The ones that stay alive usually take them seriously early. They understand how the business actually works, from engineering to buyers to cash, and they know when the founder has to change with the company.

Bernie Bulkin also gets into how founders can stay out of some of these lethal traps. I’m turning that into a separate Millennial Masters feature. In the meantime, you can get Why Start-Ups Fail here.

That’s usually where things start slipping. The pitch gets stronger, the business underneath it gets harder to explain, and one simple question can give that away very quickly.

That’s why I asked Montgomery to write this guest piece. She’s backed multiple companies that reached billion-dollar outcomes, writes the Angels and Unicorns newsletter, and has a new book coming out under the same name.

What she gets into here is one of the clearest investor tells I’ve read in a while. 👇🏻

The pitch is working 🔍 The business isn’t

The pitch was excellent. It had a tight narrative, a clear picture of the market, and the founder knew their fundraising history cold. He also had a decent answer on competition.

He was also wearing a Patagonia vest, which somehow felt on-brand for the whole thing. By the time the deck ended he’d said the word “momentum” five times. I remember because it’s one of the words I always keep a tally of when I’m being pitched.

I made one request. “Walk me through last month.”

The long pause was eventually followed by, “We’re seeing really strong traction across multiple channels and the market response has been very positive.” I stopped taking notes.

That’s not a business update, it’s a brand statement. I’ve heard a version of it in more first meetings than I care to count, usually from founders who’ve spent three months on the pitch narrative and about seven minutes on their numbers.

What I’m actually watching for

By the time I’ve agreed to meet you, I’m not really watching the deck. The deck is what got us in the room in the first place.

I’m watching for the gap between what you’re saying and what’s actually happening in the business. This is easier to spot when I’m tired and grumpy, which probably says something unflattering about late afternoon slots.

A deck is designed to close that gap. If it got us together, then it’s a good deck. You’ve worked hard on it. But the gap shows up in how you answer the question that wasn’t in the prep and that no deck ever covers.

So I make the same request in every first meeting. Not because I’m so smart, but because it works and I haven’t found anything better.

Walk me through last month

Founders who know their business can usually answer in under sixty seconds. They give me one number and what moved it, or one decision they made or didn’t.

Whether the number is good or bad is almost beside the point, which still surprises some people when I say it. What I’m looking for is operational clarity.

Founders who are running on narrative answer with narrative. Trends. Channel response. Market signals. Healthy pipeline. No number, no cause, no decision.

The longer the answer, the wider the gap. I’ve never had a founder answer like that and turn out to be running a tight business underneath it.

Maybe the exception exists somewhere. I’ve mostly stopped expecting it.

The thing I don’t know how to fix

Most founders in a fundraising process have spent more time on the pitch than on their operating data in the months leading up to it. I understand why. The pitch gets you in the room. Nobody schedules a meeting because your MRR tracking is clean.

But somewhere in the preparation, the pitch stops being a representation of the business and starts replacing it. The story becomes load-bearing and the founder starts knowing their TAM better than what happened last month.

The strange thing is that by the time they’re in front of me, they’re not exactly lying. They’ve just been in pitch mode long enough that the pitch feels like truth.

I don’t have a clean answer to this. The obvious one is “stay close to your numbers”, but founders raising a round are doing twelve things and the numbers rarely feel like the most urgent one.

What I can tell you is that the founders who stay in contact with their operating data during a raise are noticeably easier to talk to. The meeting just feels different, more relaxed and less performative.

If you’ve already closed a round

Don’t think this stops applying once you’ve raised. You’ll likely raise again.

The next round’s investors will be later stage, less patient with narrative, and quicker to spot the gap. What passes at seed doesn’t pass at Series A.

Not because Series A investors are smarter, but because they’ve sat through more of these meetings and the filter gets tighter with exposure.

Run the test on yourself now. What’s your answer? Pick last month. One number and what caused it. One decision you made as a result. Say it out loud. Time it. If you can’t do it cleanly in sixty seconds, it’s worth asking why before someone else does.

The pitch should be good. That part’s not wrong. It just shouldn’t be the thing you know best when you walk in.

👤 Susan Montgomery backed 13 unicorns at pre-seed and seed. In Angels and Unicorns, she writes about what investors are actually clocking in those first ten minutes, and why founders often miss it until it’s too late. Subscribe for a clearer view from the other side of the table.

Susan | Angel Investor

I backed 13 ‘unicorns’ at pre-seed and seed. Writing for angels who want better returns and for startup founders who want their attention.

Two more reads before your next investor pitch

I’ve been there, and it’s frustrating.

That’s why I asked executive coach and adviser to write about how to get aligned with your team and clear confusion before it turns into culture.

Kathy is a 2x CEO and COO, writes the Lead Without Limits newsletter, and works with leaders on fixing themselves before they start trying to fix everyone else. 👇🏻

Where confusion starts

When teams start missing things or doing work twice, the problem often starts earlier than you think.

The instruction sounded clear enough in the room. But once the work got moving, it turned out not to be clear at all.

That’s because capable people still need to know what’s actually being asked of them.

You thought you were clear

If you are a leader, especially early in your career, there is a bias a lot of organisations push on you: sound more confident than you feel.

I got that advice in my mid-twenties, just after stepping into my first executive role at a large media company.

My boss told me I would be more effective if I answered questions with more confidence and fewer caveats. What I heard was simple: stop sounding uncertain.

When I pushed back and said, what if I don’t know the answer, what if I’m not sure, he said to come off as sure. That’s what people expect.

What happens next is harder to notice. You get used to sounding certain, and after a while you stop checking whether people understood what you meant in the way you thought they would.

That’s where the question starts to matter. Instead of asking yourself whether you said it clearly, ask what the other person actually heard.

The team fills in the gaps

This problem doesn’t stop with the leader. It moves through the team as well.

In some teams, people learn quite quickly that asking too many questions makes them look slow or unsure. After a while, they stop asking and just try to keep up.

That’s one of the biggest gaps inside a team. People start acting as if they should already know what the leader meant and move forward without needing much else.

Something else happens as well. People learn to talk about what is going fine and stay quiet about the rest. Anything messy, unclear, or unfinished gets left unsaid.

Over time, fear starts shaping what people say and what they hold back. The team gets less honest about what is not working, and problems stay untouched for longer than they should.

Where the damage shows up

This starts in a very ordinary way. Someone gives direction too quickly, assumes it was clear enough, and moves on.

I know because I was one of those leaders. My to-do list was so long it could not be done in a day, sometimes not even in a year.

In a startup environment, it can feel like the whole organisation is sitting on your shoulders. When that pressure builds, speed feels like the only sensible response because everyone is waiting on you.

That’s when the direction goes out too quickly and some of the detail gets lost.

The team feels the rush as much as they hear the words, so people stop looking for more context and move ahead with whatever they think they understood.

Then it spreads. One unclear conversation becomes several, and by the time it reaches vendors or partners, people are already working from different versions of the same expectation.

You start seeing work repeated, and people get frustrated with each other because they don’t realise they’re working on the same thing.

After a while, trust starts to go because nobody is completely sure what was meant to happen or who was meant to do it.

You don’t need a weak team for that to happen. It can happen with good people when the expectation was never as clear as everyone thought.

Most “clarifying” makes it worse

One of the biggest mistakes you can make at this point is deciding for yourself what needs clarifying instead of finding out what the other person actually took from it.

You can try to fix the problem by explaining more, but that does not always help.

Sometimes you answer the original question, add extra detail around it, and leave people more confused than they were to begin with.

Or you can say a lot more and still leave the real point untouched. What matters is whether you dealt with the part that had actually gone wrong.

That’s why the better question is not did I say it clearly? Ask what the other person actually heard.

The other mistake is treating clarification like another announcement. It works much better when it becomes a conversation.

Support is not control

Not everyone on your team needs the same level of oversight.

A newer team member who is still learning the role usually needs more check-ins. Someone experienced and motivated can work with much lighter ones.

I think about it as an 80/20 problem. For someone who needs more support, I might stay close to roughly 80% of the work in progress. For someone stronger and more self-directed, maybe 20%.

The numbers are not really the point. The point is that different people need different levels of support, and a lot of leaders forget that.

The next part gets missed a lot. Don’t just tell people when you will check in. Ask what helps them do the work well.

People don’t all need the same thing. One person may prefer to give updates in writing, while someone else works better by talking things through.

Someone learning something new might need a daily check-in for a while, and another person may find that amount of contact suffocating.

You don’t have to follow every preference, but it helps to understand what makes the work easier for each person to do well.

What will make a difference is building a way of working that gives people the support they actually need.

Silence is not alignment

When nobody has any questions, don’t assume that means things are clear. More often, people don’t feel safe enough to say what’s not making sense yet.

If everyone walks away from an important conversation saying “that was fine” or “sounds good,” and nobody has anything they want to push on, that should worry you.

A team that’s really thinking something through will want to test at least one part of it. They might ask what happens if this goes wrong, or whether something has been missed.

The same thing applies in one-to-ones. If you want the real answer, give people a heads-up that you want their honest read on something specific, and be clear about what you are asking. Then ask and pause.

It’s easy to rush in and fill the silence too quickly. Leave it there for a moment, and that’s usually when people say what they really think.

Start where the cost is highest

Start with one or two things, not ten.

Sit down and ask where the biggest gap is right now. You don’t need the full picture straight away. Focus on the area where unclear expectations are causing the most trouble or wasting the most time.

Write it down, be honest about what is not working, and then do the part most leaders resist by naming it openly with your team.

You don’t have to have it figured out. You can say: I’m not sure I’ve been setting clear enough expectations in this area. I want to do better, and I’d like to understand what you need.

That kind of honesty doesn’t weaken your authority. It gives people room to tell you what is actually getting in the way.

Then get specific about the gap. Are you clear on the outcome but vague on how people should get there? Or have you been clear on behaviour without explaining why it matters? Start there.

Build in one way of checking this and actually use it, starting with the area that matters most and the team closest to it.

Change shows up in the work first

You’ll feel the difference before you can measure it properly.

People start saying sooner when something is unclear, and problems get picked up before they have had time to spread.

That’s the first sign things are getting clearer. The team spends less time cleaning up confusion later because more of it gets dealt with properly at the start.

👤 Kathy Wu Brady is an executive coach and adviser, a 2x CEO and 2x COO, and the writer behind Lead Without Limits. She has spent 20 years launching, scaling, and reshaping businesses across five industries, and now helps leaders make clearer decisions and lead with more courage. Connect with Kathy on LinkedIn

Lead without Limits

Find your path to meaning and success with the help of a 2x CEO & COO, executive coach, and artist who's been in the arena.

By Kathy Wu Brady

🤜🤛 More teamwork from Millennial Masters

If AI keeps giving you bland writing, obvious ideas, and work you still have to rescue, the problem usually is not the model.

You haven’t really trained it, briefed it, or given it enough context to do the job well. You’ve been throwing work at it and hoping for the best.

You wouldn’t onboard a human hire this way. AI is no different.

To get real value from it, you need to treat AI more like staff and take the time to train it properly. That’s the point Millennial Masters guests keep coming back to.

A better question is what a new hire would need to know in order to do this well. Ask it that way and the mistake becomes obvious. 👇🏻

You skipped the onboarding

Carly Meyers explained it in the clearest way I’ve heard: “Think of it like one person. Think of it like you’ve just hired someone.

“Say you just hired a sales and marketing person. They’ve walked into your office for the first time. You sit them down and say, right, this is what I want you to do.”

That’s the part most founders skip. They don’t explain the business properly or spell out who the customer is.

They also don’t show the AI what the brand sounds like, what counts as good work, what needs approval, or where it should stay in its lane.

They start with tasks, which is the same mistake people make with new hires when they are moving too fast. Things get messy quickly.

That’s not new. Strong hires still need context, and good employees take time to become useful. Throw someone into the middle of the business without proper onboarding and the work usually comes back half-baked.

Then AI gets treated as if none of that applies. The expectation is close to finished output on day one, and when it disappoints, the tool gets the blame.

This is a management problem

Damon Flowers was direct: “AI is like a staff member and it can do anything. It can do finances, it can do marketing, it can have a conversation with you and negotiate stuff.

“But you’ve got to be really good at training it around who it is, what’s the role, what context you’re giving it, and what do you specifically want it to go do for you.”

That’s the shift a lot of founders still haven’t made.

If AI is sitting inside your workflow like a member of staff, the output depends on the job you have given it and how clearly you have set it up.

If that setup is vague, the output usually comes back vague too. You get writing, research, strategy, and sales copy that could belong to almost anyone.

Jess Jensen made the same point: “I think of AI as almost like a junior employee. The more that I train and invest in that junior employee to get them to understand our strategy, our goals, our vision, who I am, the better the output’s going to be because now they’re really tracking with me.”

That’s a much better approach than treating AI like a one-shot tool. The quality improves when it starts following how you think and the standards you work to.

Open a fresh window, type a request, get something back, and judge the whole experience on that one exchange. Of course the output will disappoint.

There’s no real context, no memory worth using, no examples from previous work, and no proper explanation of how the business operates.

Ryan Carruthers described AI as “a marketing or copywriting assistant right now… like having a graduate, someone who’s just getting into it, who’s starting to learn the frameworks.”

That’s useful because it brings the expectation back down to earth. A graduate can help and save you time. They can also waste a lot of time if you expect them to understand the business from one instruction and a few scraps of documentation.

You’re expecting senior output from something you barely trained, then blaming the tool when it cannot deliver.

Give it a job first

Start by getting clear on the role. A lot of AI use is still far too broad.

If you ask it to be a strategist, researcher, editor, analyst, and operator all at once, it ends up being none of them properly. If you want better work, define exactly what role it is supposed to play in the business.

That might mean a copy assistant, a research analyst, a sales support function, a coding reviewer, or an editorial second pair of eyes. The more precise the role, the easier it becomes to train well.

Carly Meyers’ framing helps here. One “person” does one type of work. The AI is no longer guessing what it is there to do.

The practical test is simple. If a human joined your team tomorrow and got this exact brief, would they know what role they were in and what good looked like in that role? If the answer is no, the AI will struggle too.

Once the role is clear, context starts doing most of the heavy lifting. A lot of people cut corners here because giving context feels slower than prompting. Skip it and the quality usually falls apart.

Nick Holzherr described AI as “a distributed team member,” which is useful because anyone working outside the room needs clear documentation to do the job properly. Nobody expects remote employees to just pick all of that up on their own.

It has to be written down somewhere, with examples and enough detail for someone to understand how the company operates. AI is no different.

That’s why businesses with decent documentation and defined processes often get stronger results from AI sooner. The material needed to train it already exists.

If everything is still sitting in your own head, the output is usually worse because the AI is being asked to work inside a business that has never been clearly articulated.

Teach it what good looks like

You get weak output when you expect AI to pick up voice and standards from nothing. Give it a clear sense of how the business sounds at its best.

That means showing it strong examples of how you write, sell, and communicate, along with the kinds of phrasing or claims you do not want it using.

This matters because voice is not decoration. In a lot of businesses, it carries judgement. It tells you whether something feels salesy, technical, or generic. Leave that undefined and the model usually fills the gap with average internet sludge it’s been trained on.

Yota Trom uses AI as an editor rather than a creator, saying it helps her because she’s not a native speaker. That’s a useful boundary. It gives the AI a clear job and makes it easier to see what should still stay with the human.

The same thing applies to the goal of the task. If you ask for writing, research, summaries, or strategy thinking without explaining what the work is supposed to achieve, the AI has to guess what good looks like, and the output usually gets weaker.

When the task has a clear purpose, the output usually improves. AI is no different.

Constraints matter too. If it cannot make claims without sources, say that. If it should not sound too polished, say that. If it needs to protect margin, stay within brand language, or avoid legal overreach, make that clear as well.

You don’t need a perfect instruction sheet for every task. You need to stop expecting AI to know the business better than the person using it.

This is where it gets useful

Don’t ignore this part because it is less exciting than trying a new tool.

The quality improves when you start correcting the work and letting that shape the next round. That is how a useful working relationship develops, whether it is a junior hire or a custom GPT.

Jess Jensen’s point about investing in the junior employee matters here. It takes repetition, correction, and enough patience to build something better over time instead of opening a fresh session and starting the whole disappointment cycle again.

What looks like the model getting smarter is often just better management. That is where the gains stop being theoretical.

Once the system is trained properly, more first-pass work gets handled before a human steps in. Drafts come together faster, research is easier to pull together, and more of the admin-heavy work gets moved off people’s plates.

James Augustin thinks about AI automation as taking the admin-heavy parts of a role away so people can focus on the work they are actually good at and paid for.

Ryan Carruthers gave a similar example when he described building an AI that could handle after-hours sales work, gather the right information, and book calls so nobody on the human sales team had to keep carrying that burden late into the evening.

Damon Flowers sees it as getting work to 85 to 95% done before a person even touches it. That number will vary by task, but the principle is right.

At that point, you are reviewing and refining far more often than you are starting from scratch. That’s where the time starts coming back.

The real mistake you are making

AI is starting to expose which businesses are systemised and which still live mostly in the founder’s head.

Weak output comes from shoddy setup. The role is vague, the context is thin, the standards live in your head, and then the tool gets blamed for sounding half-baked.

You’re expecting too much from something you’ve barely trained.

That’s why this is more fixable than you think.

You don’t need a better model to get better results. You need to brief it clearly, train it properly, give it feedback, and hold the standard.

Do that and you’ll get more from AI than most people complaining about it ever will.

I first heard it mentioned in one of my early Millennial Masters interviews with Yehong Zhu, who referenced Nassim Taleb’s Antifragile in her book recommendations.

By then, I had already lived through years where stability disappeared.

The pandemic period, especially 2020 through 2022, was one of the hardest stretches I’ve had in business and in life.

I had seen the shocks, the uncertainty, and plans breaking down without having a useful language for it, let alone a framework for how to respond.

That is why ’s take stood out. When we spoke about antifragility and I read how he translated it into the day-to-day reality of running a company, I knew it would make a strong guest post for Millennial Masters.

His piece below turns antifragility into something far more useful than theory. It is the practical version of what founders can do when the old playbook stops working. 👇🏻

How you build for shocks

It is easy to get used to a stretch where demand is steady, costs are predictable, and suppliers pick up the phone.

That has not been the world for a while now. Tariffs shift, energy prices move, and supplier pricing changes without much warning.

A Harvard survey of more than 4,000 SMBs found that nearly half of owners misjudged how tariffs were affecting their costs.

That is not because small operators are careless. It is because the world around them is changing faster than most people can track while still trying to run a business day to day.

Last October, I published a detailed piece on Operating translating Nassim Taleb’s ideas on antifragility into business practice. The framework still holds up, but frameworks do not pay invoices.

This is the practical version. It is for the founder with 12 employees, one line of credit, and a Monday morning stand-up with very little patience for theory.

By antifragile, I mean a company that can take a shock and come out stronger on the other side. The goal is to build in a way that leaves you better placed when things move against you.

The 5 buffers that keep you in the game

Buffers are what keep a bad quarter from becoming a crisis.

If they are too thin, one shock can put you on the back foot. If they are strong enough, you have time to respond and, sometimes, enough room to take advantage.

1️⃣ Cash

How many weeks can you operate if revenue drops 30 percent tomorrow? Not based on a trailing twelve-month average, but on your actual weekly burn against your actual cash balance.

If the answer is less than eight weeks, that deserves your attention before almost anything else. When I was scaling R.G. Brewton from $27 million to north of $75 million in industrial distribution, the metric that gave us permission to be aggressive was knowing exactly how many weeks of runway we had.

2️⃣ Time

Slow decisions kill small companies faster than most bad decisions do. How many days does it take you to reprice a product, switch a supplier, or shut down a project that is bleeding money?

If the answer is weeks, you have a problem. Tariff policy changed eleven times in six months last year. The founders who got through it repriced in 48 hours. The ones who waited absorbed the margin hit.

3️⃣ Pipeline

If your top three clients represent more than 50% of revenue, your P&L can look steadier than it really is. It holds until one of them leaves, delays payment, or renegotiates terms.

Concentration risk is one of the most common buffer gaps I see in advisory work. It does not show up on a dashboard. It shows up on a Tuesday morning phone call.

4️⃣ Talent

Name every person in your company who is a single point of failure for a critical process. If that person gets sick, quits, or burns out, what breaks?

That is a structural weakness. Cross-train the role or document the process. Do that this month.

5️⃣ Suppliers

How many of your key inputs come from one source, one country, or one contract with no pre-vetted alternative?

In B2B distribution, we kept backup supplier relationships for every critical SKU category. Not because we assumed we would need them, but because if we did, there would be no time to start from scratch.

Small bets and fast feedback loops

Taleb’s barbell strategy is simple. Keep 80 to 90% of resources in proven work that pays the bills. Put 10 to 20% into small bets with real upside. The danger sits in the middle: the mid-budget initiative with medium conviction and no kill switch.

For a small company, a small bet is not something you talk about on a strategy retreat. It might be a landing page with $300 in ad spend behind a new offer, a pilot engagement with one client in a new vertical, or a price test on a single product line for two weeks.

What matters is that the hypothesis is written down before you spend the money, the downside is capped, and the timeline is short enough to read before the next move.

The kill rule matters more than the bet itself. Define what failure looks like before you start, not after sunk cost fallacy has you giving it another month. If a test cannot show signal within two to four weeks, it is either too slow or being measured badly.

Review active bets once a week. Spend fifteen minutes looking at what is still open, what should be killed, and what is showing enough signal to push harder.

Done properly, this is how small wins start to stack without small losses quietly piling up underneath them.

Guardrails to stop a bad month from spiralling

One bad month is manageable. The problem starts when a second bad month passes with no real response and a third is allowed to land on top of it. What matters most is how quickly you react.

This comes down to urgency, both yours and your team’s.

Three triggers should force a same-week decision:

• Cash drops below your minimum weeks-of-coverage threshold

• A top-three client signals churn or payment delay

• The cost of a key input spikes by more than 15% with no clear date for it to come back down

When one of those triggers fires, do not leave it hanging. Make the call that week. The response might be small. Reprice one line, pause one project, or call one client. What matters is that you move.

Run a subtraction audit once per quarter. List every recurring cost, meeting, subscription, and initiative, then cut the bottom 10 to 15%. This is not about being tidy. It is about reducing the number of places where a shock can hurt you.

Every unnecessary cost gives rising prices another place to squeeze margin. Every unnecessary meeting takes time away from fixing what matters when something breaks.

Before any commitment above a number that matters to your business, spend ten minutes writing down how it could fail and whether the structure of the bet protects you if it does. It is Taleb’s premortem at founder scale.

It takes ten minutes. It has saved me from more bad decisions than any financial model.

5 items. 10 minutes. Weekly.

That is the entire system. No single review will change much on its own. The value comes from doing it 52 weeks in a row.

The forecast is dead. Your business shouldn’t be.

You cannot run a company as if conditions will stay still long enough for the spreadsheet to catch up.

What matters more is whether the business can absorb pressure, respond quickly, and adjust when something changes.

This is practical, not theoretical. The founders who keep moving well under pressure are usually the ones who have already built for it.

👤 John Brewton writes Operating, a bestselling newsletter documenting the history and future of operating companies. He graduated Harvard University and began his career as a PhD student in economics at the University of Chicago. After selling his family’s B2B industrial distribution company in 2021, he spent the years since helping business owners, founders, and investors improve how they operate. He’s the founder of 6A East Partners, a research and advisory firm focused on the future of companies. He still cringes at his early LinkedIn posts and loves making content every day, despite protests from his wife, Fabiola. Connect with John on LinkedIn.

Operating by John Brewton

Weekly Reporting on the History & Future of Operating Companies

I spent years watching startups burn through funding, trying to out-build incumbents, and I made the same mistake on products I worked on.

The teams that lasted focused on less.

That lesson is why I asked to come back and break down how lean teams build competitive advantage.

In his last guest post, he broke down how to find gaps in a competitor’s strategy. This time, he walks through what to do once you find them. 👇🏻

The edge comes from precision

The trap most teams fall into with AI

Building an impressive demo is easier than ever with AI.

Competitors can spin up tools in a weekend that look great on the surface. But those screens rarely integrate with a customer’s existing setup, so they fall apart in real production.

Big competitors optimise for speed and broad appeal. Lean teams win by spotting small friction points those tools ignore and building around them.

AI design platform Magic Patterns reached $1M in ARR with no team. Co-founders Alexander Danilowicz and Teddy Ni built and scaled it on their own, managing to raise a $6M Series A.

Tools like v0, Bolt, and Lovable generate interfaces from prompts, but the output often feels generic and disconnected from the actual product.

Danilowicz and Ni came from front-end development, which shaped how they approached the problem.

They focused on design system integration, letting teams import their own components, so prototypes match the existing product from the start.

That focus created an advantage broad tools struggle to replicate without changing how they work.

How lean teams build moats

Saying no matters

Lean teams work within constraints, which forces focus.

Magic Patterns said no to customers who wanted backend support. They focused on product teams of 5–50 people working with existing design systems.

Each decision made the positioning clearer. Product teams with design systems created better expansion and retention than one-off project work.

A customer might start with a couple of licences and expand across the team as the product proves itself. You see the same pattern elsewhere.

In the AI meeting space, Granola made a similar decision. They ignored the mass transcription market and focused on C-level executives and VCs in sensitive meetings.

Both companies built multi-million-dollar businesses by committing to very specific customers, which shaped everything else they built.

Where the advantage actually comes from

Magic Patterns focused on understanding what product teams were trying to build and preserving that in the output.

The Chrome extension pulls structure from existing sites and reuses components so outputs match the product.

That keeps them strong in a specific part of front-end work where generic tools struggle.

Changing what each role can do

Speed doesn’t hold for long, so advantage comes from somewhere else.

Magic Patterns shifts what different roles can do inside a team. When Zeal’s CTO Pranab Krishnan started using it, he stopped writing long UX descriptions and began building interactive prototypes himself.

Engineers no longer wait on design handoffs, and they build with a clear sense of what can actually ship. Even product managers can lay out requirements visually. The product team can do more without adding new roles.

That kind of shift makes a tool hard to remove once a team adopts it. I break down the full strategy of Magic Patterns in more detail in Product Strategy Decoded.

The lean team competitive advantage audit

Three quick tests you can run on your product:

1️⃣ The rejection test

Who are you explicitly saying “no” to? If your product tries to serve very different customers at once, the focus disappears.

Pick the customer you understand best and build for them. Every customer you turn away sharpens how well you serve the ones you keep.

2️⃣ The workflow test

Does your product fit into how people already work? Magic Patterns integrated into the design systems teams already use.

AI tools that ignore existing context are easy to replace. The closer you are to real workflows, the harder it is to replace you.

3️⃣ The capability test

Are you only speeding things up, or changing who can do the work? Tools that only save time compete on shrinking margins.

Tools that let non-specialists contribute become part of how teams work. Does your product let new roles do work they couldn’t before?

The real advantage

Building advantage as a small team comes down to focus, not feature volume.

It comes from a clear understanding of the customer and helping them do things they couldn’t before.

Every market leader operates on a set of assumptions. Over time, those assumptions stop matching reality.

That’s where lean teams find their opening. They spot constraints larger competitors ignore and build for the customers who feel that gap most.

The competition becomes useful if you pay attention.

Start with one question: what are the biggest players in your market not set up to solve? That gap is where your advantage sits.

👤 Mike Goitein writes Product Strategy Decoded, where he reverse-engineers the strategy of products, startups, and creators using a proven framework. He’s a Certified Strategy Professional with over 15 years of experience helping Fortune 200 companies and startups alike sharpen their choices and deliver measurable results. Connect with Mike on LinkedIn.

Product Strategy Decoded

Every week, I pull back the curtain and make strategy easier using a proven framework. You'll get insights of what works today and why, with powerful takeaways you can apply to your own business.

By Mike Goitein

Over the last two years I’ve lost a significant amount of weight using GLP-1 medications. They definitely work. The part nobody explains comes after.

I’ve tried all three of the big ones along the way: Ozempic, Wegovy and Mounjaro. I now stay on a small maintenance dose because the main thing people say about them is actually true.

Weight-loss injections quiet the constant thinking about eating. When that noise fades, a surprising amount of attention comes back to the work you actually care about. Anyone running a company knows how valuable that mental space can be.

The part people talk about less is what the process feels like while the weight is coming off. These drugs work by pushing your calorie intake down automatically.

When that happens your body runs on less fuel, which means energy can dip, mood can wobble a bit and exercise is usually the first thing that disappears from the week.

That’s usually the moment you realise the injection handles the appetite, while everything else still depends on you. You feel lighter on the scale while your body sometimes feels worse.

Another practical point is cost. Unless you have them prescribed, GLP-1s can become a serious monthly expense (up to £400). When you are spending that kind of money, it makes sense to think about how to get the best outcome from them instead of hoping the injection will solve everything on its own.

And sooner or later most people face the same question: what happens when you stop or reduce the dose, because the food noise returns and the habits underneath start running the show again.

That’s why I asked Tom Hutchinson-Smith about this.

Tom is a personal trainer I interviewed previously on Millennial Masters and he works mostly with founders and operators whose schedules look messy from the outside but still want their body to keep up with the demands of running a business.

I asked him to explain how GLP-1s actually fit into that reality and how training and nutrition keep the process from weakening the system underneath. 👇🏻

Why the injection suddenly makes sense

Founder life rarely supports neat routines around food.

The day fills with calls, travel, meetings and deadlines. Meals happen wherever they fit and hunger tends to appear at inconvenient moments.

Over time eating becomes something that constantly interrupts the day. GLP-1 medications change that environment quickly.

Hunger fades into the background and the urge to snack stops driving the schedule.

Meals shrink naturally and eating becomes something you do when needed rather than something you think about all day.

For someone already juggling a company that change feels like relief.

What the scale doesn’t show you

GLP-1 drugs make eating less much easier to maintain. That part works very well.

What they don’t do automatically is protect strength or muscle while bodyweight falls.

When calorie intake drops for months at a time the body adjusts. Muscle can shrink along with fat and strength in the gym slowly slips.

The scale continues moving in the direction people want, so this often goes unnoticed.

For founders that matters more than it might seem. Your body is the machine that carries long days, travel, pressure and constant decision making.

When strength fades the work feels heavier even though nothing in the calendar has changed.

The training that keeps your body intact

Tom’s advice to founders is deliberately simple because complicated plans rarely survive busy schedules.

Three strength sessions during the week keep muscle active while weight comes down. Each session fits inside an hour and focuses on movements that train large parts of the body at once.

The real focus is keeping the engine running properly while the rest of the business keeps moving.

When founders travel frequently Tom usually gives them a shorter hotel-gym version of the same routine so the habit doesn’t disappear during busy weeks.

Tom shared the exact framework he gives founders using GLP-1s. 👇🏻

Download The Lean Leader Program and use it as a practical guide.

What to eat when you barely feel hungry

When GLP-1s suppress appetite the amount of food drops quickly. That means the food you do eat needs to carry more of the workload.

Protein helps hold on to muscle during the weight loss phase. Vegetables help digestion and keep the body getting nutrients even when meals are smaller.

Dietary fat still matters for hormones and general health even though extremely heavy meals tend to feel uncomfortable on these medications.

Tom usually encourages founders to think less about strict dieting and more about building meals around protein and vegetables first.

This approach usually covers what the body needs without turning food into another project to manage.

The body that runs your business

GLP-1 medications are powerful tools and they genuinely help many people lose weight.

They remove the constant battle with appetite which used to make fat loss extremely difficult to sustain.

What they don’t replace is the physical foundation that keeps someone functioning well while running a business.

Strength training, regular movement and basic nutrition habits protect the body that supports your work. When those stay in place the weight loss process leaves someone lighter without leaving them weaker.

That outcome usually feels very different once the drug is no longer doing the work.

The number on the scale changes quickly. The condition of the body running the company matters far longer.

Discipline still decides the outcome

One part of the conversation that rarely gets mentioned is what happens if you stop taking the drugs.

From my own experience the food noise does not come back overnight, yet after about two weeks the difference becomes noticeable.

Hunger creeps back in, thoughts about snacks start appearing again, and that quiet mental space around food begins to shrink.

That is the reason I now stay on a very small maintenance dose. I do not always follow the seven-day schedule either. Sometimes it stretches to ten days, sometimes fourteen, usually depending on when I feel that background noise returning.

Keeping the result still comes down to discipline.

GLP-1 medications make the early stages of weight loss far easier to sustain, yet keeping the result afterwards depends on the habits that support your body when the drug is no longer doing the work.

If those habits disappear, the body gradually returns to old patterns.

The people who keep the result are usually the ones who built the habits while the drug was helping them.

Because once the noise returns, the discipline is yours.

Quick question for founders reading this:

If you’ve tried GLP-1 injections, what surprised you most about the experience?

I’m curious how others are navigating this.

More healthy stuff from Millennial Masters:

But that’s not usually how the useful stuff works inside a business.

The tool is there. The output is weak. The instinct is to blame the model.

makes a better case than most for why that usually misses the point.

She runs Build to Launch, where 5,000+ builders learn to ship real products with AI.

Her guest piece gets into what changes when you treat AI less like a chatbot and more like a hire. 👇🏻

The hire nobody trained you for

Most founders interact with AI the way they’d interact with Google: type something in, hope for the best, get frustrated when the output is mediocre.

That’s like hiring someone talented and never telling them what the company does.

AI is the teammate who never sleeps, has read everything on the internet, works at impossible speed, and has zero judgment about what actually matters to your business.

Left unmanaged, it produces confident-sounding work that misses the point entirely. Managed well, it becomes the most reliable operator on your team.

The shift isn’t technical, but managerial. And if you’ve ever onboarded a brilliant but unpredictable new hire, you already have the instincts for this.

Set direction before you set tasks

When a new team member joins, you don’t hand them a task on day one and expect perfect output. You give them context: what does the company do, what does quality look like here, what are the unspoken rules. AI needs the same onboarding.

In my workflow, I maintain what’s essentially an employee handbook for AI. It’s a document that spells out: who we are, what our voice sounds like, what standards we hold, what mistakes to avoid, and what “done well” looks like. Every single task the AI touches starts from that foundation.

When I skip this step (and I still do sometimes, because speed is tempting), the output is generic. It reads like it could’ve been written for anyone. The moment I set direction first, the output sounds like it came from someone who actually works here.

The management principle: Context before tasks. Always.

If you’re a founder delegating to AI for the first time, write down three things before you open any tool:

1. What does your company/brand actually sound like? (Not “professional.” Be specific. Conversational? Data-driven? Blunt?)

2. What are the non-negotiable standards? (Accuracy over speed? SEO matters? No jargon?)

3. What should AI never do on its own? (Publish without review? Make claims about customers? Touch financials?)

That document becomes your home truth. Write it once, and every interaction with AI improves from that point forward.

Give feedback, don’t start over

Here’s the pattern I see constantly: someone gives AI a task, gets a mediocre result, scraps it, and tries a completely new prompt from scratch.

That’s the equivalent of firing someone after their first draft.

The founders getting real value from AI treat bad output as a feedback opportunity. When the first result misses the mark, they tell it specifically what’s wrong.

Not: “This isn’t good, try again.”

Instead: “The tone is too formal for our audience. Make it conversational, like you’re explaining this to a smart friend over coffee. Also, the second section jumps to a conclusion without enough evidence. Add a specific example.”

That’s a performance review. And it works exactly the way it works with people: specific, actionable feedback produces better results than vague disapproval.

I routinely go three to five rounds of feedback on a single piece of work. Not because AI is bad at its job, but because good output is iterative.

The same is true when you’re managing humans. The difference is AI doesn’t get defensive, doesn’t take it personally, and can turn around revisions in seconds.

The management principle: Iterate, don’t restart. Your feedback is the most valuable input in the system.

Consistency from systems, not supervision

If you’ve ever managed a growing team, you know the moment when things start breaking: when the work depends on you being in the room.

The fix is always the same. You create systems: templates, checklists, standard operating procedures. AI works identically.

I run 15+ specialized AI configurations for different types of work: research, writing, code review, SEO, editing. Each one has its own “job description” with specific instructions, quality standards, and guardrails.

The research agent knows to cite sources. The editing agent knows our style guide. The code review agent knows our security standards.

This means I don’t re-explain the basics every time I start a task. The system holds the standards. I just point it at the work.

For founders and operators, this is the unlock. You don’t need 15 agents. Start with one reusable prompt template for the task you do most often:

• If you write weekly updates, create a template that includes your voice, your format, your typical structure.

• If you review proposals, create a checklist prompt that evaluates against your actual criteria.

• If you do market research, create a research brief that specifies what sources you trust and what depth you need.

Save it, reuse it, and refine it over time. That’s how you scale AI output without scaling your own involvement.

The management principle: Build systems once. Run them repeatedly. Improve them incrementally.

Know when to step in & when to let it run

The hardest management skill, with people or AI, is knowing when to get involved.

Here’s the decision framework I use, and it works the same way any experienced manager delegates:

Let it run when the cost of a mistake is low: internal notes, first drafts, brainstorming, data formatting, meeting summaries. If an error takes 2 minutes to fix, the review isn’t worth your time.

Review before it ships when the output is client-facing, public, or financially significant: proposals, published content, anything with numbers that someone will act on. AI is confident even when it’s wrong. Your job is to catch the difference.

Do it yourself when the task requires your specific judgment, relationships, or lived experience: strategic decisions, sensitive communications, anything where your reputation is on the line in a way that a quick fix can’t undo.

Most founders err in one of two directions. Some review everything, which defeats the purpose of having a fast teammate. Others review nothing, which eventually produces an expensive mistake they could’ve caught in 30 seconds.

The management principle: Delegate based on the cost of failure, not the difficulty of the task.

The job has changed, and that’s the opportunity

Here’s what I want to leave you with.

The founders getting disproportionate results from AI often realised that their role shifted. Their focus moved from doing the work themselves to defining what good work looks like and shaping how it gets done.

That’s a management role. It involves setting direction, giving feedback, building systems for consistency, and deciding where output can run with autonomy and where it needs closer review.

The tools will keep getting better and the models will keep improving. What compounds over time is your ability to manage the process and shape how the work flows. That skill carries forward as the systems evolve.

You’re developing a management capability that scales with each new AI system that ships.

If you’re reading this thinking you already know how to manage, that foundation still applies here. It simply needs to be directed at this layer of work.

Start here

Everything in this article came from doing this hands-on, not reading about it.

If the management mindset clicked for you and you want to see what it looks like in practice, I wrote a companion piece that walks through 12 projects in a deliberate progression: from a 10-minute build to a system that runs your content, research, and publishing without you touching it.

If you’re a founder or operator, skip straight to Tier 3. That’s where the management principles from this article become real infrastructure: a personal command center that encodes your best workflows, a research pipeline that runs autonomously, and a one-center ecosystem where you sit in one place and everything amplifies outward. That’s the part built for people who think in systems, not code.

And if you want this kind of thinking every week with practical AI systems, real examples, that’s what Build to Launch is. 5,000+ founders and builders getting the playbook for making AI actually work.

👤 Jenny Ouyang runs Build to Launch, a newsletter and community for builders shipping real products with AI. She’s built and launched 10+ products, published 100+ articles, and created the systems described here, where AI operates as a managed team, not a magic wand. Subscribe to Build to Launch.

Build to Launch

Turn AI ideas into shipped products. Proven systems and showcase opportunities for builders who launch.

By Jenny Ouyang