Executive Talent Magazine
While executive search firms race to adopt AI for efficiency gains, Russell Reynolds Associates is taking a radically different approach: building what their new Chief Science Officer calls a “synthetic brain,” a proprietary OpenAI-powered system that ingests proprietary data and insights to fundamentally transform how leadership potential is identified and predicted.
Dr. Tomas Chamorro-Premuzic is an international authority in people analytics, and the Chief Science Officer at Russell Reynolds Associates, a founding member of the Association of Executive Search and Leadership Consultants (AESC).
From Individual Expertise to Collective Intelligence
Where advantage in the profession once came from who you knew, what matters now is what your organization knows. Dr. Chamorro-Premuzic is trying to solve this paradox: brilliant individual consultants whose collective knowledge is inaccessible and untapped.
“What I’m trying to do is transfer individual knowledge into a collective brain that makes the firm smart as a system,” he explains. “Individually, people are very smart and experienced and have a lot of expertise, but collectively, it doesn’t necessarily systematize or transfer.”
The competitive landscape has fundamentally shifted. A decade ago, a consultant might win an engagement by personally knowing every viable CFO candidate in a region. Today, clients can ask ChatGPT to scrape the Internet for candidate profiles, but clients can only access experience, expertise, judgment, and trust through people.
Building the Synthetic Brain
“We just started an exclusive partnership in the sector with OpenAI,” Chamorro-Premuzic reveals. The firm’s version of GPT is connected to market intelligence, assessment information, and more. However, allowing an algorithm to access data is not enough. The challenge lies in making decades of accumulated knowledge usable.
“There is a lot of data in our systems, and my initial job is to make it usable from a legal, ethical, and data science perspective,” he notes. “The main thing is to always have a human in the loop.”
Chamorro-Premuzic is quick to confirm the goal is augmented judgment. In his view, even with ideal data and perfect tools, it would be unthinkable to have an autonomous system decide on executive appointments without human intervention at multiple levels.
Revealing Bias
As subjective data from thousands of anonymized conversations populates RRA’s synthetic brain, a critical question emerges: how do you prevent institutionalizing bias? Chamorro-Premuzic says, “The first thing is to truly capture everything, and then before the information goes from being an insight to being an action, we run it past the scrutiny of what we call ethical AI algorithms, that identify existing biases,” he explains.
For instance, if analysis reveals that positive candidate references skew heavily male, that gap gets flagged. According to Chamorro-Premuzic, the science shows gender differences in leadership are either non-existent or favor women. Identifying the disconnect between what clients want and what the data shows is step one. Step two is persuading clients to act on the science rather than their preferences.
“To then persuade our clients that what they want is not what they need is the art of influence and persuasion,” says Chamorro-Premuzic. “I can tell all of my consultants, guess what? Certain qualities of leaders indicate greater overall performance – have you thought of this unusual profile? They still have to do their best to persuade them… I think a lot of clients can be persuaded by data.”
He says, “The worst bias of all is to blame AI algorithms for being biased when, in fact, they’re actually exposing our own biases.”
The Formula for Leadership Potential
What signals actually predict leadership success? Chamorro-Premuzic offers a formula. “If I’m looking at a recipe, I would say 20 or 30% is going to be what you have done in the past,” he begins. Past experience is necessary but not sufficient. “It’s a hygiene factor that gets you in the door.”
The remaining 80% breaks down as follows:
- 20% cognitive capacity: “I think weirdly in our field and in the industry, in general, intelligence defined as mental horsepower and learning ability is actually underrated,” he notes.
- 20% drive and ambition: “Not being pathologically ambitious and greedy, but being motivated by the right things, which is to succeed through others and bring value.” This bucket also includes integrity, increasingly critical as scandals demonstrate that narcissistic dispositions predict eventual self-destruction.
- 40% people skills: “Can you motivate? Can you energize? Can you inspire? AI can do the IQ part of things, but the EQ part… if you’re a really technical candidate, but you [lack EQ], you’re going to need a lot of coaching.” This category also encompasses coachability, the capacity to grow.
This framework deliberately elevates people skills to 40% of the equation, recognizing that as AI handles more analytical work, the distinctly human elements of leadership become more valuable, not less.
What Clients Want from AI
AESC’s most recent research captures what clients, the purchasers of executive search and leadership advisory services, want and expect from AI. The findings are clear: clients want efficiency gains in lower-level tasks including market mapping, document preparation, and administrative work, while insisting on human judgment at the decision level.
“I think they are right,” says Chamorro-Premuzic. “The closest we are to autonomous hiring or autonomous talent identification is AI and algorithms working at the low end of the spectrum. I spent seven years at ManPower Group doing this, and there it makes total sense because you have 50,000 candidates that you have to find in two weeks… You can make a lot of individual mistakes. But if you’re appointing a CEO, you cannot make a mistake. The stakes are so much higher, and the costs are… catastrophic.”
The metaphor he offers is apt: dating apps work fine for volume matching, but if you’re selecting a prince for a princess, you need all the tools plus a committee.
Breaking the “Looking Under a Lamppost” Pattern
Perhaps the most persistent challenge in executive search is what Chamorro-Premuzic calls the looking-under-the-lamppost problem: boards searching for talent only in familiar places, not because that’s where the best candidates are, but because it’s the only place they know to look. It’s the parable of searching under the lamppost—not because you lost your keys there, but because that’s where the light is.
“The same companies and boards that complain that there’s talent scarcity and they can’t find the right person are looking for talent in the same old places,” he observes.
Science-based assessment offers a way out: evaluating candidates based on leadership potential rather than traditional credentials. With AI amplifying pattern recognition, firms can identify candidates who’ve never worked in a specific industry but possess the cognitive capacity, drive, and people skills to excel there.
The Path Forward
Dr. Chamorro-Premuzic is making a bet that the future of the industry lies in systematizing expertise without losing the judgment that makes it valuable.
“We have more and more clients that are saying, okay, I’m going to think outside the box and not go for the usual suspects,” Chamorro-Premuzic notes. “You’re telling me this person has never worked in consumer or private equity or tech, but they could do it because they have the profile of somebody who can drive forward and execute the strategy.”
For AESC Member firms, the implications are profound. The competitive question is no longer “Who do you know?” but “What does your organization know, and how effectively can you deploy that knowledge?” Individual brilliance must evolve into collective intelligence. Data must be transformed into defensible insights. And science must augment, not replace, the human judgment that remains irreplaceable at the highest level.
The firms that solve this puzzle will define what comes next for executive search and leadership consulting. In an age of artificial intelligence, competitive advantage belongs to those who can make human judgment scalable.
Dr. Tomas Chamorro-Premuzic is the Chief Science Officer at Russell Reynolds Associates. As the firm’s first Chief Science Officer, he leads R&D, innovation, and data strategy, with a focus on using science and data to inform leadership advisory services.
He holds positions as a Professor of Business Psychology at University College London and Columbia University. He has also been the CEO of Hogan Assessment Systems and Chief Innovation Officer at ManpowerGroup. Dr. Chamorro-Premuzic is based between London and New York.
