The Danger of AI Advice Without Commercial Challenge

AI is already changing how founders and business owners think, plan and make decisions, and I do not see that as a bad thing.

Used well, AI can be incredibly useful. It can help structure thinking, test ideas, prepare options, summarise information and speed up some of the work that sits around decision-making.

I use it myself as part of how I prepare, reflect and improve my own practice, not as a substitute for commercial judgement or professional experience, but as a tool that can support better preparation and sharper thinking when it is used with care.

What I am also seeing, however, is another side to it.

More founders are now bringing AI-generated thinking into mentoring conversations. Sometimes it is a strategy, sometimes a marketing plan, sometimes a funding narrative, sometimes a set of leadership actions or operational improvements. On the surface, much of it looks impressive because it is structured, confident and often persuasive, but the problem is that confidence is not the same as judgement.

AI can produce a lot of useful material very quickly, but it does not truly understand the business in front of it. It does not know the real constraints, the personalities, the cash pressure, the history behind previous decisions, the customer dynamics, the founder’s energy levels, or the trade-offs sitting beneath the words. It can create a convincing answer, but it cannot always judge whether that answer is commercially right for this business, at this stage, with these people and these resources.

That is where the risk begins, particularly for founders, because the heavier the decision, the more dangerous it becomes to mistake a confident answer for a commercially sound one.

AI can take you deeper without making you clearer

One of the biggest dangers I am seeing is not that AI always gives a bad answer, but that it gives a lot of confident answers.

A founder starts with a sensible question, follows the thread, asks for more detail and explores another angle, the AI agrees with each stage, reinforces how promising the thinking looks, how valued the ideas are and before long the founder is looking at a twenty-point plan, a new framework, three possible strategies, a revised proposition and a long list of things to do. It feels productive because there is output, but output is not the same as clarity.

The founder may not have made a better decision. They may simply have gone further down a rabbit hole, surrounded by seemingly polished thinking that has not been properly tested against the reality of the business.

That matters because most founders I meet are not short of ideas. More often, the real pressure comes from having too many competing options, limited time, limited cash, limited management capacity and a lack of clear prioritisation. In that situation, more thinking does not always help, especially if it creates another layer of noise rather than a clearer decision.

For a founder-led business, this can become a real problem because the issue is rarely a lack of possible actions. It is knowing which action deserves commitment, which decision needs to be made now, and which attractive-looking option should be left alone.

The real value sits in the challenge

A good AI answer can still need a commercial sanity check.

The question is often not whether the answer sounds sensible, but whether it is relevant to the business now, what assumptions sit behind it, what it may have ignored, what would have to be true for it to work, and what it would cost in cash, time, energy and attention. Just as importantly, someone needs to ask what would receive less focus if this became the priority.

These are not theoretical questions. They are the questions that turn information into judgement.

AI can suggest a strategy, but it cannot take responsibility for the commercial consequences of pursuing it. It can produce a growth plan, but it cannot properly judge whether the business has the leadership capacity, cash headroom or operational strength to execute it. It can make something sound coherent, but it cannot always tell whether the founder is solving the right problem.

That is where professional business mentoring creates real value, not by rejecting AI, but by challenging, interpreting and grounding the output so that it becomes useful rather than simply impressive.

The mentor is not there to dismiss the work the founder has done with AI. In many cases, that work is a useful starting point. The value is in helping the founder test it against the commercial reality of the business, sharpen the decision and move from broad possibility into focused action.

The business mentor’s role becomes more important, not less

I do not see AI as a threat to good business mentoring. If anything, I think it makes high-quality mentoring more important because when a founder can generate more information than ever before, the differentiator is no longer access to information, it is the quality of judgement around that information.

The value is not simply in knowing what could be done. It is in deciding what should be done, in what order, with what level of commitment, and with a proper understanding of the trade-offs involved.

That is where an experienced business mentor, grounded in real commercial experience rather than mentoring technique alone, helps the founder step back from the volume of content and reconnect with the commercial reality of the business. My work with founders is often at the point where the decisions have become heavier, the trade-offs are less obvious and the cost of getting it wrong is higher. The issue then becomes whether the proposed course of action strengthens the business model, improves margin, supports cash generation, builds leadership capacity or contributes to long-term value, rather than simply creating more activity and complexity.

These questions are important because the effects of these decisions are not isolated. A new market affects focus, a pricing change affects cash, a senior hire affects culture and cost, a funding decision affects control, risk and future options, and a new opportunity inevitably affects everything else already competing for attention.

This is why a polished answer can still be the wrong answer.

The danger is not AI. The danger is unchallenged AI.

The issue is not that founders are using AI, in many cases, they should be.

The issue is using it as if it has the final word.

That can create false clarity because the output sounds credible, so the thinking feels validated. The plan looks detailed, so the decision feels more advanced than it really is. The strategy sounds intelligent, but may still be wrong for the stage, resources or direction of the business.

Good business mentoring interrupts that pattern by helping the founder slow down, test the assumptions and separate what is interesting from what is important. It brings context, commercial judgement and challenge into the conversation, and it helps distinguish between a useful idea and a priority worth committing to.

That is difficult for any system to replicate because good mentoring is not just about the answer. It is about reading the person, the business, the moment and the decision in front of them. Sometimes the right intervention is challenge, sometimes it is simplification, sometimes it is encouragement, and sometimes it is helping the founder stop before they commit time, cash and energy to something that sounds better than it really is.

And that last point really matters. Founders are often very good at creating momentum, but momentum in the wrong direction can become expensive. The value is not simply to keep moving, it is to create the right momentum around the right decisions.

AI should improve the conversation, not replace it

For me, the strongest use of AI in business mentoring is not to replace the commercial conversation, but to improve the quality of preparation so the conversation can go deeper.

A founder can use AI to organise their thinking, explore options, prepare scenarios or draft an initial plan, and that can be very useful. The real value, however, comes when that thinking is brought into a properly challenging conversation and tested against the business.

That is where the mentor can help clarify the actual decision, test the commercial logic, examine the evidence, expose the assumptions, understand the risk, shape the sequence and identify the action that will create meaningful progress.

This is where AI and mentoring can work well together. AI can help create the raw material, but business mentoring helps turn that material into better judgement, clearer decisions and stronger execution.

Because in business, the issue is rarely whether something sounds intelligent. The issue is whether it improves the decision.

My straight view

I am in favour of AI, and it will of course become more and more part of how many founders think, plan and prepare. Used well, it is a powerful tool.

But I would be cautious about any business owner relying on AI advice without challenge, especially where the decision affects cash, people, funding, strategy or future direction. AI can help you think faster, but faster thinking is not always better thinking, and it can sometimes take you much further into a decision before anyone has properly asked whether it is the right one.

The value of an experienced business mentor is not to compete with AI or pretend that AI does not exist. The value is to bring commercial judgement, perspective and challenge to the output, so the founder does not mistake volume for clarity, confidence for correctness, or activity for progress.

My straight view is this: AI may help you produce more answers, but a good business mentor helps you make better decisions, especially when the decisions carry real weight.

For a growing business, that is where better judgement becomes better execution.

About the author

Mark O’Neil is the founder of Kinetic Mentoring and works with founders and leadership teams on strategic decisions around growth, focus, funding and execution.

Mentor to Founders When the Decisions Get Heavy.

Clarity. Momentum. Results.