It is so easy to mistake velocity for actual progress. Buying a new tool takes five minutes, but fixing a messy, broken workflow takes real effort.
The real skill moving forward isn't adopting every piece of software that drops...it's having the discipline to cut out the noise and keep things simple. Subtraction is highly underrated.
Just creating activity doesn’t create value. With building becoming easier than ever before, the differentiation will come from knowing what to build, rather than just build
The gap you don’t notice in the demo but really feel a few months in is the part nobody budgets for, and it tends to fall on the people furthest from the decision to buy. In the regulated businesses I work with, the real sticking point is who actually holds the authority to switch a shiny tool off once it’s costing more than it saves. Does anyone properly own that call in these organisations, or does it slip between teams?
The maintenance cost is one of the sneaky parts. A tool can become wildly expensive once people are hunting through Slack to work out who owns the mess.
Adoption telling you who clicked while absorption shows what people actually trust is a sharp distinction.
Appreciate that, John. Absorption > Adoption. Thats how matured AI programs scale.
It is so easy to mistake velocity for actual progress. Buying a new tool takes five minutes, but fixing a messy, broken workflow takes real effort.
The real skill moving forward isn't adopting every piece of software that drops...it's having the discipline to cut out the noise and keep things simple. Subtraction is highly underrated.
Just creating activity doesn’t create value. With building becoming easier than ever before, the differentiation will come from knowing what to build, rather than just build
Appreciate the comment!
Without clear systems and human judgment, complexity can quietly drain energy and focus from teams.
One extra check here, one “can you review the AI version?” there, and suddenly the time-saving tool has become another colleague who needs managing.
Precisely the point. Appreciate your comment.
I've found documentation matters more than prompts.
If your team can't explain how the work gets done, AI can't reliably improve it. It just follows whatever process you hand it, good or bad.
Prompts get all the romance because they feel like a secret spell. 🪄 Documentation is where the ghosts live. 👻
I couldn’t agree more. You’ve got to simplify stuff. Prompts aren’t magic.
It is so true!!
Appreciate that.
The gap you don’t notice in the demo but really feel a few months in is the part nobody budgets for, and it tends to fall on the people furthest from the decision to buy. In the regulated businesses I work with, the real sticking point is who actually holds the authority to switch a shiny tool off once it’s costing more than it saves. Does anyone properly own that call in these organisations, or does it slip between teams?
Every shiny tool has a buyer. Very few have a designated adult with permission to say, “This thing is now expensive furniture.”
That is spot on. From my work with fintech and financial sponsors, I have a similar view. It’s the easiest thing to plan for, but often overlooked.
More tools often amplify existing process issues rather than solving them
The maintenance cost is one of the sneaky parts. A tool can become wildly expensive once people are hunting through Slack to work out who owns the mess.
Indeed. The engineering and BAU processes required to maintain the slop far outweigh the benefits.
If it doesn’t deserve a human to write it, it doesn’t deserve a human to read it!
Point well made!
Yes, and Neha’s line on subtraction is where this gets uncomfortable.
Everyone wants AI to add capacity. Fewer people want to admit the process itself shouldn’t survive contact with a half-decent automation review.
That is spot on. Simplify before you can automate. The business case for subtraction is very compelling.