In an age where anyone can sound like an expert just by asking AI, how do you tell the real experts apart? Harvard Business Review has been asking this very question throughout early 2026. And the answer is surprisingly simple — it's not the people who talk that survive, it's the people who actually do.
What is this about?
In the March 2026 issue of HBR, John Winsor (Executive Fellow at Harvard Business School) published a provocatively titled piece: "Has AI Ended Thought Leadership?" — asking whether AI has killed thought leadership as we know it. His diagnosis goes like this: thanks to generative AI, the cost of looking like an expert has dropped to essentially zero. Anyone can ask ChatGPT and get a plausible-sounding insight.
The problem is that organizations are drowning in this "plausibility." Polished slides, well-crafted reports, impressive keynotes — they're everywhere, but almost none of it leads to actual change. Winsor calls this "expertise theater". And the barrier to entry for this theater has hit zero.
So his proposed alternative is the "Thought Doer". Not someone who advises from a safe distance, but someone who builds pilots with the team, shares failures openly, and takes responsibility for outcomes. His argument: this is the only way experts can prove themselves in the AI era.
5 Red Flags of a Fake Expert (HBR)
1. No battle scars from failure — Can't describe specific failures in detail
2. Altitude Lock — Can't get down to ground-level details
3. Generic failure language — Just repeats "I learned from failure" with no concrete lessons
4. Expertise built too fast — Time spent posturing as an expert exceeds time actually doing the work
5. Views never change — Tech landscape shifts dramatically but their positions haven't budged in 2 years
This isn't just Winsor's take. If you look at the AI leadership articles HBR published between January and March 2026, there's one common message — "When AI democratizes knowledge, what's left is judgment and execution".
What's actually changing?
The core shift comes down to three things.
First, judgment has become the new scarce resource. According to David Duncan's piece in HBR, AI dramatically boosts the productivity of experienced professionals, but it can actually be harmful for juniors. Because AI handles all the tedious, repetitive work, the very opportunities where judgment used to develop simply disappear. Duncan calls this the "apprenticeship crisis of judgment".
Second, a new role is emerging: the "agent manager." According to a co-authored piece by Harvard Business School professor Suraj Srinivasan and Salesforce COO Vivienne Wei, as AI agents get deployed in real operations, a dedicated role to manage them has become necessary. Just as product managers emerged during the software revolution, agent managers are predicted to become an essential role of the AI revolution.
Third, it's people — not technology — that are the bottleneck. In HBR's annual survey, 93% of respondents said "the key barrier to AI adoption is culture and change management". Only 7% blamed the technology itself. That's an all-time high.
| Old "Expert" Model | AI-Era "Expert" Model | |
|---|---|---|
| Proving expertise | Knowledge volume, tenure, credentials | Hands-on experience, failure cases, prototypes |
| Leader's role | Decision-maker, director | Agent orchestrator, judgment coach |
| Developing juniors | Natural learning through repetitive tasks | Intentionally designed stretch experiences |
| Key bottleneck | Technical skills, information access | Judgment, change management, culture |
| Value differentiator | What you know | What you've done + contextual awareness |
A Harvard Business School study analyzing 50,032 global software developers backs this up. Developers using AI spent 5% more time coding and 10% less time on project management. Team sizes also shrank. In other words, AI is reducing the very need for middle management. If you're a manager, you need to shift your weight from "work I manage" to "work I personally create" to survive.
"As AI absorbs analytical tasks, the source of differentiation shifts to human judgment, insight, and the ability to build meaningful relationships. These cannot be automated or accelerated."
— David Fubini, HBS Senior Lecturer
The essentials: Capabilities leaders need in the AI era
- Become a "Thought Doer"
Instead of keynotes and reports, run an 8-week embedded sprint and build a pilot yourself. Share your failure stories with the team. Winsor's advice: "Ask for failed experiments. Not polished case studies." - Design a judgment development system
When AI takes over repetitive work, juniors lose their natural path to building judgment. Like medicine or the military, you need to intentionally design case-based learning, simulations, and gradually expanding responsibilities. - Introduce the agent manager role in your organization
You need a dedicated role that monitors AI agent performance metrics (quality, speed, escalation rates) and optimizes prompts and workflows. HBR predicts this will become a standard job title within 12–18 months. - Build Change Fitness
Advice from HBS professor Tsedal Neeley: at least 30% of your entire workforce needs digital/AI literacy. You have to move three layers simultaneously — individual curiosity, team-level collaboration patterns, and organization-level data infrastructure. - Protect the "meaning of work"
A warning from HBS professor Jon Jachimowicz: if AI boosts productivity by 20% but meaning drops by 20%, the net effect is zero. When a customer service rep loses the chance to talk directly with customers, motivation vanishes too. When automating, design for "meaning" alongside "efficiency."
Watch out: The paradox of AI-era leadership
Here's the paradox the HBR survey data reveals.
- Companies that say AI investment is their "top priority": 99%
- Companies that say they're getting "high or significant business value" from AI: 54%
- Percentage citing "culture and change management, not technology" as the key barrier: 93%
The technology is ready, but organizations can't keep up. The leader's role has shifted from "technology adoption" to "organizational change design."
