How to find the people rebuilding their work around AI
On most teams, the loudest voice about AI and the most valuable person using it are rarely the same human.
One demos a new tool in every standup. They ask for more licenses, forward the latest model announcement, and bring AI up in every meeting. Theyâre easy to spot, which is exactly why they end up with the training budget and the leadership attention.
The other person says almost nothing. They quietly turned a two-day task into a four-hour one and moved on to the next thing. Youâd never know unless you went looking at what they actually shipped.
When the time comes to scale AI across the team, most leads fund the first person. Thatâs the expensive mistake.
Enthusiasm is loud. Output is quiet.
The numbers make the gap obvious. One analysis of more than 400 companies found AI tool usage climbed 65 percent while median pull request throughput rose only about 8 percent. A lot of motion, very little progress.
That disconnect is the whole story. Most teams measure who talks about AI, not who has been changed by it. Training flows toward the vocal users. Budget follows enthusiasm. The people who genuinely rebuilt their workflows stay invisible the entire time, because real change rarely announces itself in a standup.
A vocal user adopts the tool. A real contributor redesigns the work around it. Only one of those compounds when you scale it.
Output reveals what behavior hides
Youâll find the signal in the work itself. Standups wonât hand it to you.
Pull the last quarter of output and look for the curves that bent. Find the person whose turnaround time dropped with no loss in quality, whose revision counts keep falling, who now ships the same volume in half the hours or double the volume in the same time. That person changed something structural about how the work happens.
Someone who compressed a multi-day process into an afternoon has built something you can study and reuse. Someone who gives a polished demo has built a presentation. For a scaling strategy, the first is worth ten of the second.
Interview the workflow itself
Once youâve found them, resist the urge to ask which tools they like. That question gets you a product review and nothing more.
Ask how they actually do the work now. Walk me through the last time you ran this end to end. Where does AI come in, and where do you still work by hand? What did you stop doing completely, and what broke before you got it right?
Youâre after the sequence, the judgment calls, and the small corrections theyâve stopped noticing they make. Most people who have rebuilt a workflow canât explain it cold, because it has turned into instinct. Your job is to slow them down until the instinct becomes steps again. Record the conversation. The throwaway comments are usually where the real method hides.
Extract it before it lives in one head forever
A redesigned workflow that exists only in one personâs head is a liability. They take two weeks off and the speed leaves with them.
Document it while youâre still in the room. Capture the prompts they reuse word for word, the order they work in, the checks they run before anything ships, and the moments where they override what the AI suggests. Write down the failure modes too, because half the value sits in what they have learned not to trust. Keep all of it concrete enough that someone else could follow it next Monday without coming back with a single question.
Turn one personâs process into a system
A personal workflow becomes leverage the moment it survives a handoff.
Give it to one other person first and watch where they get stuck. Their confusion will expose the steps the original owner skipped, the ones so obvious to them they forgot anyone would need them spelled out. Fix those gaps, then widen the circle. Fold the prompts into shared templates. Wire the quality checks into your existing review process so standards hold steady as more people run the same play.
Do this well and the four-hour workflow quietly becomes the team default. The person who built it finally gets credit for the thing that moved the needle, rather than the colleague who demoed the most tools.
Thatâs how AI workflows actually scale. You find the quiet ones who already cracked it, then make their work repeatable for everyone else. Your most valuable AI user has probably already transformed how they work and never mentioned it. Look at what shipped, and youâll see them.
. . .
Want to save hours each week by turning work into repeatable AI workflows?
The Fortune 100 AI Skills Library⢠includes plug-and-play prompts built to save leaders time and money. Copy, paste, and edit in 60 seconds, then apply them across planning, execution, and reporting.
Top comments (0)