Deliberate Practice

I recently had the chance to learn about improvisation with Axel van Exter, a world champion. In this theatrical art, the rules are simple and ruthless: the audience proposes a setting or a line, the performers have seconds to pick up the cue, build on each other’s ideas, and land a coherent scene.

What separates champions from participants is the ability to recognize a pattern as it forms and respond to it before the moment passes. The best improvisers hold a deep library of previously encountered configurations such as types of characters or narrative turns built through years of repetitions or ‘reps’ in practice and performance. When a familiar configuration appears, effective moves present themselves quickly while ineffective ones are discarded almost without conscious thought.

In chess, research has established that grandmasters proceed in similar fashion. They store tens of thousands of meaningful patterns or ‘chunks’ in long‑term memory. These templates are recognized instantaneously, allowing them to make their moves under time pressure.

Whether on stage, in chess, or during live high-pressure events such as earnings calls, M&A negotiations, panel discussions, interviews or debates, participants do not have the luxury to review all their options with considered pros and cons. Instead, they must scan for patterns based on the recognition-primed decision process developed by psychologist Gary Klein: They identify the situation, generate a plausible response, mentally test it, and act, relying on experience-based intuition.

The ability to operate at the highest level under pressure must be earned. Psychologist Anders Ericsson described the engine behind expert performance as ‘deliberate practice’: highly structured, effortful work aimed at improving specific skills, with constant feedback. Over reps typically spanning a decade or more, experience is compressed into patterns, and patterns are compressed into reflexes. Critically, he showed that deliberate practice can extend performance well beyond what talent alone would predict.

Artificial intelligence (AI) relies on a form of systematic pattern recognition, too. During training, AI learns how to match new inputs to vast numbers of stored data structures and generate likely sequences. In fact, AI already matches or exceeds human performance in closed games such as chess and in generating plausible language on the fly. It can even support live improvisation as a co‑performer.

Pattern recognition has its limitations. However similar a new situation may appear to a previously encountered one, it is never exactly the same in open, uncertain settings. In that context, judgment is required to adapt the pattern-induced solution to specific circumstances.

It follows that, if AI industrializes pattern recognition, human judgment will remain paramount when operating with responsibility under uncertainty. Expertise built through deliberate practice remains crucial.

There is a second derivative. Using AI, in itself, requires expertise and judgment. Only deliberate practice with reps and active learning will allow both to develop. The instrument is new. The path to mastery is not.