Never Start a Team at Zero

By Martin Bergmann | AI Project Lab | September 2026

Three people on this team used AI this month: one to synthesize customer feedback into themes, one to draft a first pass at a design option, and one to catch a gap in a requirements doc that a manual review missed. Nobody counted any of it.

The uncounted head start.

It’s easy to write this off as not real progress. Scattered, unofficial, nobody coordinating it, it doesn’t feel like the same thing as a team actually being ready. That is not nothing. That is real progress. It just hasn’t been counted.

Why counting it matters more than it sounds like it should.

In 2006, researchers Joseph Nunes and Xavier Drèze ran a loyalty-card study at a car wash. One group received a card that required eight stamps to earn a free wash. Another group got a card requiring ten stamps, but two were already stamped; the same eight purchases still stood between them and the reward. 19% of the first group finished their cards. 34% of the second group finished theirs, despite needing to do exactly the same amount of work. Progress that’s visible, even progress you didn’t have to earn, changes how likely people are to finish.

This is not the same claim as the one behind the Generation Effect post a few weeks back, and it’s worth saying that directly rather than leaving readers to wonder. That piece was about how people build a skill: struggle first, see the better version second, or the learning doesn’t stick. This is about something else entirely, whether people finish something they’ve already started. A head start doesn’t teach anyone anything. It just changes whether they keep going. Handing a team a finished prompt library still kills the learning. Telling a team they’re starting from nothing when they aren’t still kills the momentum. Different mechanisms, different failures, and confusing the two is the actual mistake.

Where AI actually helps.

Try this before you launch anything formal: every approved project already has a charter or a business case; that's standard practice, not a special requirement. Point AI at that, then at what the organization already has: a similar project last year that already tested this vendor, a different team's market research that answers half of this project's open questions, and the retro notes from the last attempt at something like this. Ask it to find what's actually relevant. Nobody has time to connect a new project to years of scattered organizational work by hand. AI does, in minutes.

Bring that back to the team before day one of anything formal. Not "welcome, we're starting from scratch." Closer to: here's where you already are.

Nobody starts a team at zero. Most people just haven't been shown where they actually stand.

---

Martin Bergmann is the Director of an IT PMO and founder of the AI Project Lab. He writes about AI, project management, and the future of intelligent work.