AI Doesn't Level the Field. It Amplifies Who You Already Are.

In Harvard Business Review, economists Christopher Stanton and Simon Rudat and Terret CEO Justin Shriber share new research in What Top Performers Do Differently with AI, and Why They See the Biggest Benefits. They interviewed eight chief revenue executives, met with more than a dozen sales leaders and surveyed 236 sales professionals at Stripe and Checkr. Their conclusion pushes back on a hopeful assumption. AI helps most workers, but in these leading-edge firms the biggest gains are going to the top performers.

The reason is not simply that top performers save more time, although they do. After adopting AI, managers gave higher-value work to 44% of top-quartile reps, compared with 12% of bottom-quartile reps. The deeper difference is how the best people use the tools. They treat AI as a thinking partner across the whole deal, to sharpen strategy, rehearse negotiations and learn from past conversations. The authors call them autodidacts: people who learn by experimenting, comparing notes and iterating without waiting for a training session.

The Advantage Is How You Learn

The study measures AI literacy, and each one-point rise on its five-point scale was linked to a 0.71-point rise in the benefits people reported. Trust in AI, by contrast, had little to do with the gains. What sets the winners apart is knowing how they learn, and using the tool to extend strengths they already understand.

One rep at Checkr built a negotiation copilot before such tools were widely available. A Brex employee built a tool over a weekend because it seemed useful, and the company later rolled it out. Nobody assigned those projects. They came from people who knew their strengths and went looking for ways to use them more.

I read this research as a story about identity as much as technology. The authors conclude that AI gives the most lift to people who can build systems around their natural strengths. You can't do that if you have never been asked what your strengths are. Many professionals have spent years editing themselves to fit a job description, doing the work the way the process prescribes. AI rewards the opposite habit. It amplifies the person who knows what they naturally solve for, and it does little for the person who is still waiting to be told.

None of this is fixed. What you solve for keeps evolving, and curiosity can be learned. The environment decides a great deal of it.

Autodidacts Are Grown, Not Just Hired

When the researchers asked managers what drove success, the answers were about the room more than the individual: executives who model AI use, explicit encouragement to experiment, peers sharing what works and concrete examples from power users. In other words, self-directed learning thrives where it is visible, permitted and shared. In a culture that punishes mistakes or rewards only the approved workflow, even naturally curious people learn to keep their experiments to themselves.

The authors make a useful hiring point. Instead of asking candidates whether they use AI, ask how they learn. One leader they interviewed asks candidates to "share with me the AI session you used to prepare for this interview." I like that question because it shows how a person thinks.

Their entry-level example may be the most important part of the piece. Brex could have fully automated its sales development role. Instead, it automated 20 to 40% of the manual work and let ambitious reps use the freed time to close full deals on their own, with a path to promotion. The redesigned role attracted stronger talent and increased internal promotions. That is a company protecting the first rung of the ladder while raising its ceiling. I wrote about what happens when that rung weakens in Confidence Is Leaving Through the Entry Level.

The risk in this research is a two-tier workforce, where the people already ahead pull further away while everyone else is handed a script. The answer is to give everyone the chance to discover their own edge. Structure still matters. Clean data, shared definitions and guardrails let experiments become reliable tools the whole team can use. The authors call it centralizing context and decentralizing discovery, and that balance is what lets one person's breakthrough lift everyone.

Questions Worth Asking This Week

For leaders: who on your team has built something with AI that nobody asked for? Find them, ask them to show their peers, and make that kind of sharing a regular habit.

For managers: in your next one-on-one, ask each person what part of their work they are best at and how AI could help them do more of it. Many people have never been asked the first half of that question.

For individuals: name the problem you naturally solve better than most, then spend an hour this week testing how AI could extend it. As I argued in The Return of the Generalist Starts With Knowing What You Solve For, the people who keep learning across boundaries are the ones the AI era will need most.

AI is magnifying a difference that was already there, between people who know who they are and people still waiting for permission to find out.

The more powerful our tools become, the more they reward people who know what they bring.

Want to explore these ideas further? Learn more about my work on leadership, identity and conviction at www.theglennllopis.com.