- In a study of 5,179 customer-support agents by Brynjolfsson, Li and Raymond, an AI assistant raised output by 14% on average, but 34% for the least experienced staff and little for the most experienced. The tool spread the best workers' methods to everyone, closing the gap between novice and expert faster than training alone.
- That flips a familiar assumption. AI's biggest lift often lands on your newest people, not your senior ones, which changes how you hire, train and value experience.
- The mechanism is knowledge transfer. The AI carried the habits of the top performers and handed them to everyone else, so newer staff reached experienced-level performance quickly.
- Your experts still matter, differently. Their edge on routine work narrows, while their value on the hardest cases and on teaching the machine what good looks like rises.
- Handled well, this builds a flatter, faster team: new people contributing sooner, experienced people freed for the problems only they can solve.
You watched your newest hire close a hard case before lunch. Six months ago that same case would have taken them a week and two escalations. Now the AI is at their elbow, and they handled it cleanly. Then you glanced at your most experienced person, the one you quietly built the team around, and their numbers had barely moved. Nothing is wrong. But the shape of your team is rearranging under you, and it is worth understanding why.
Here is the reframe. Most leaders assume AI rewards their best people most, because the best people use everything best. The evidence points the other way. The largest gains are landing on the least experienced, and that single fact changes how you build a team. It is good news, if you see it early and plan for it.
Who actually gains most from AI at work?
The newest people, by a clear margin, at least in the strongest study we have. Researchers Erik Brynjolfsson, Danielle Li and Lindsey Raymond followed 5,179 customer-support agents as an AI assistant was rolled out, and published the work in the Quarterly Journal of Economics. On average, the tool lifted output by 14%. For novice and lower-skilled agents it lifted output by 34%. For the most experienced agents, the effect was small.
Read the scope before you generalise. This is one company, one type of work, thousands of agents. It is a strong signal, not a universal law. But the pattern is striking and it fits what leaders describe across many settings: the tool lifts the floor much more than the ceiling. It is the practical face of the idea that the advantage lives in knowledge, not in the tool itself.
Why does AI help beginners more than experts?
Because of what the tool actually does. It had learned from the best agents, and it handed their habits, phrasings and judgement calls to everyone else in real time. A beginner with that support behaves more like a veteran, sooner. An expert already had those habits, so there was less for the tool to add. The AI compressed the experience curve: newer people reached strong performance faster than years on the job would have taken.
That is worth sitting with, because it quietly changes the economics of a team. If a capable newcomer reaches good performance in months rather than years, then hiring, onboarding and the value you place on raw experience all shift. The lower-altitude question is which AI tool to buy for the contact centre. The better one is this: if the floor rises this much, how should we build the whole team? It is the same move as shifting your people from doing the task to designing and running it.
AI often lifts your newest people most, not your best. It hands the veterans' habits to the beginners. That changes how you hire, train and value experience.
| What the study found (Brynjolfsson, Li & Raymond, QJE) | Figure |
|---|---|
| Customer-support agents studied | 5,179 |
| Average lift in issues resolved per hour | 14% |
| Lift for novice and lower-skilled agents | 34% |
| Lift for the most experienced agents | small |
So how should a leader build the team around this?
You stop treating experience as the only ladder, and start designing for a floor that has risen. The goal is to capture the lift on your newer people while pointing your experts at what only they can do. Work it in this order:
- Let new people take on more, sooner. With good AI support, capable newcomers can handle work that once needed years. Design the role to let them, with the right guardrails.
- Point your experts at the hard edge. Move your most experienced people toward the complex cases, the judgement calls and the exceptions the tool cannot handle. That is where their value now concentrates.
- Capture what your best people know. The tool works because it learned from them. Keep that knowledge flowing into it and into your newer staff, deliberately.
- Rethink hiring and pay for a compressed curve. If people reach strong performance faster, your assumptions about experience, onboarding and progression need to move with it.
- Protect your experts' standing. As the floor rises, make sure your senior people feel valued for the harder work, not overlooked because the gap narrowed.
Redesigning a team around what AI actually lifts?
The Strategy Session works on the shape of your team in an AI-assisted world: where to let new people run, where to point your experts, and how to capture the knowledge the tools depend on. We build the plan on evidence, then move.
Book your Strategy SessionWhat does this make possible?
A team that is both flatter and stronger than the one you have now. New people contribute early and grow fast, because the tool lifts them and your culture lets them run. Your most experienced people spend their days on the problems that truly need them, not the routine that used to fill the queue. The gap between green and seasoned narrows, and the whole team moves faster for it. That is a more capable operation than one that merely bought the tools and changed nothing underneath.
Picture it a year out. Your newest hires are doing work that used to take years to reach, and they are proud of it. Your veterans are solving the cases that only they can, and they feel it. You read the shift early and built for it, so the rising floor lifted the whole team rather than unsettling it. That is what work looks like after business-as-usual, when a leader plans from how AI really lands.
Frequently asked questions
Does AI help experienced or inexperienced workers more?
Why does AI boost beginners more than experts?
How should managers respond to AI lifting junior staff?
About the author
British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.