AI can predict. Entrepreneurs still have to judge

Imagine two entrepreneurs considering the same business idea. Both ask an AI tool to identify customers, analyse competitors, recommend a price, forecast demand, and produce a launch plan. Within minutes, each receives an impressively detailed response.

The first entrepreneur treats the output as an answer. The second treats it as a collection of assumptions. That difference is entrepreneurial judgement.

As generative AI becomes embedded in entrepreneurship, founders can access more information, generate more alternatives, and complete analytical work faster than ever.

But having more analysis is not the same as knowing what to do. In fact, when polished answers are available instantly, the psychological ability to question, interpret, and act on those answers may become one of the entrepreneur's most important advantages.

What is entrepreneurial judgement?

Entrepreneurial judgement is often described as decision-making under uncertainty. But according to Dr Tony Nguyen, Interim Senior Program Manager, Postgraduate Programs at The Business School, RMIT University Vietnam, this definition needs some unpacking.

Some decisions involve risk. The possible outcomes are known, and their probabilities can at least be estimated. An established retailer deciding how much inventory to order can examine previous sales, seasonal patterns, and customer demand. The calculation may be imperfect, but useful historical evidence exists.

But Dr Nguyen said entrepreneurs frequently face something more difficult: genuine uncertainty. Under uncertainty, the entrepreneur may not know all the possible outcomes, much less their probabilities. Customers may not understand a product that does not yet exist. A new technology may change how people behave. Competitors may respond unpredictably. Regulations, social expectations, and market categories may evolve as the venture develops.

“Faced with uncertainty, entrepreneurs will have to decide which future is worth pursuing and begin directing their attention, effort, and resources toward it,” Dr Nguyen said. “Good judgement combines knowledge, imagination, interpretation, self-awareness, and the willingness to act without the comfort of certainty.”

AI makes answers easier, but not necessarily better

AI is particularly effective when a problem contains recognisable patterns, abundant data, and relatively stable objectives. It can compare options, summarise evidence, uncover correlations, identify inconsistencies, and generate scenarios at a speed no unaided entrepreneur can match. This can make entrepreneurs more capable.

In a controlled experiment involving 124 entrepreneurs, a group of researchers based in Italy and the US found that AI assistance increased the number of opportunities participants recognised. It also improved the depth of their assessments, exploitation plans, and contextual analysis.

However, the study found an important trade-off: AI use also reduced novelty in opportunity recognition and innovation in proposed exploitation strategies.

The strongest performance came from entrepreneurs who combined AI with substantial sector knowledge. Their expertise allowed them to benefit from structured analysis without simply accepting it. They could recognise when an output was useful, when it lacked context, and when experience suggested an unconventional alternative. Without sufficient knowledge, an entrepreneur may be unable to distinguish a genuinely insightful AI recommendation from one that is merely fluent, generic, or inappropriate.

Judgement is not the rejection of AI

Dr Giang Hoang, a senior lecturer in Management at RMIT University Vietnam believes the answer is not to avoid AI or romanticise unaided human intuition.

“Human intuition remains vulnerable to overconfidence, confirmation bias, and sunk-cost fallacies. AI can help expose these weaknesses by challenging assumptions and examining alternative scenarios,” he said.

According to Dr Giang, the healthiest relationship is dialogue, not delegation. Entrepreneurs can strengthen judgement by:

• Separating prediction from judgement: Use AI for repeatable, data-rich patterns, and rely on human judgement for unprecedented situations.
• Asking AI to create disagreement: Prompt the tool to generate competing interpretations and challenge assumptions.
• Protecting firsthand knowledge: Maintain direct contact with customers, observations, and operational experience.
• Running small experiments: Use prototypes, pilot programs, and pre-orders to transform uncertainty into evidence.
• Keeping accountability human: Retain personal responsibility for financial losses, team communication, and ethical consequences.

“A polished pitch or a convincing strategy can be generated in minutes. What remains difficult is deciding which assumptions deserve belief, which possibilities are worth pursuing, when the evidence is sufficient to act, and when an apparently sensible recommendation should be ignored,” Dr Nguyen added.

“These are judgements, not calculations. The entrepreneurs who thrive in the age of AI will be those who understand which questions remain theirs to answer.”

A version of this article first appeared on Psychology Today.