Adaptive Expertise

This concept unlocks the rest of Talent Management. Before this, you likely saw expertise as a thing people either have or don't. This page changes that. It separates two kinds of expertise — routine and adaptive — and shows why, in an AI-shaped world, only one of them survives. That shift underpins every later topic: why competency frameworks describe behaviour rather than knowledge alone, why learning agility predicts promotion better than past performance, what succession planning is really protecting against, and how the VRIO model explains why a person who only does things the known way can never give your organisation a lasting advantage. If you skip this page, the rest of Talent Management reads as a collection of tools; with it, you see the reason the tools exist.

Imagine two people on the same surgical team. The first has performed the same routine appendectomy a thousand times. She is fast, accurate, and her hands know the steps before her conscious mind does. The second has done the same surgery many times too, but when a complication she has never seen appears — an unusual anatomical variation, a reaction to anaesthesia no textbook predicted — she pauses, reconsiders what she assumed, and improvises a new approach that still saves the patient. Both are experts. But they are expert in different ways. The first is a routine expert: she has mastered the standard procedure so thoroughly that she executes it flawlessly, as long as the situation matches the familiar pattern. The second is an adaptive expert: she has the same deep foundation, but she can bend it to fit a situation she has never met before. The distinction matters because modern AI is automating the first kind of expertise at astonishing speed. A large language model can write a standard contract clause; a vision model can inspect a familiar defect; a speech model can transcribe a routine call. What models cannot do — at least not yet — is the second kind: noticing that the situation is genuinely new, questioning the assumptions a routine would make, and inventing a response that was not in the training data. So when people say 'AI will replace experts', what they actually mean is 'AI will replace routine experts'. Adaptive experts are not replaced; they are the ones who decide which routines to automate, which to abandon, and which new routines to build. The capacity to adapt is not a personality trait you are born with. It is a skill that develops through deliberate practice, through exposure to novel problems, and through the habit of reflecting on experience. That is the core of this page: expertise is not a destination you reach and then stop; it is a process you keep running.

Think of a chess grandmaster. They have spent years studying classic openings, endgame patterns, and tactical motifs. Their memory holds thousands of positions, and their pattern-recognition is so fast that they often 'see' the right move in seconds, without conscious calculation. That is routine expertise: impressive, fast, and reliable. But a grandmaster also faces positions that do not match any classic pattern — unusual pawn structures, material imbalances no book covers, an opponent playing a bizarre line. In those moments, the grandmaster does not search for the closest remembered pattern and force it onto the board. They slow down, break the position into underlying principles — king safety, piece activity, pawn structure — and reason from those principles to a move that has never been played before. That is adaptive expertise. Now push the analogy to its limit: this is exactly what happened when IBM's Deep Blue defeated world champion Garry Kasparov in 1997. Deep Blue was a routine expert of the highest order — it could evaluate millions of positions per second and had a massive opening book. But it could not adapt. Its strategy was fixed before the game; it could not sense that Kasparov was deliberately steering it into unfamiliar territory and change its plan mid-game. Kasparov lost that match, but it was close, and the lesson stands: a machine can beat a human at the routine game, but the human's adaptive flexibility — reading the opponent, changing style, inventing surprises — is what kept the match competitive. The analogy breaks down in one direction: chess is a closed world with finite rules, so eventually machines may cover every position. Real work, medicine, law, management — is an open world with unbounded novelty, so adaptive expertise has no upper limit. The grandmaster's skill at sensing when the pattern does not fit is exactly the skill you need to develop.

Adaptive expertise is the ability to apply deep knowledge and skill flexibly to novel, unfamiliar situations — as distinct from routine expertise, which applies that same knowledge efficiently only to familiar situations.

Your library already carries the building blocks for this idea, even though you have not studied adaptive expertise as its own topic yet. Competency Frameworks tells you that competencies are the products of systematic development rather than innate gifts; adaptive expertise is one of those developable competencies, and it is the one that matters most in an AI-rich world. Learning Agility is a close neighbour — it is the willingness and ability to learn from experience and apply the lessons in unfamiliar situations — and adaptive expertise is what that learning produces: the flexible application itself. Career Adaptability is about navigating transitions in your own work life; adaptive expertise is the same flexibility applied to the content of the work itself. On the organisational side, Dynamic Capabilities describes how firms sense change and reconfigure resources; adaptive expertise is the individual-level version of that sensing and reconfiguring. And Succession Planning is risk management applied to people — the risk it manages is precisely the loss of adaptive experts whose knowledge cannot be replaced by a procedure or a model. Only a few of these neighbours are named in the page that follows, so this map is your own.