Josh Cote, entrepreneur, consultant, and founder of EAW Consulting
AI has become the first place many people turn when a difficult business problem lands in front of them. A client stops replying, a hiring decision becomes complicated, pricing suddenly feels uncertain or a strategy problem refuses to resolve itself, and the instinct is to open ChatGPT or Claude and start typing.
At EAW Consulting, there is no mystery about why that happens. The tools are quick, always available and easy to use. Nobody has to arrange a meeting, explain the background twice or worry that the question sounds basic. For founders and leaders who are used to making decisions under pressure, that convenience is enormously attractive.
There is nothing inherently wrong with using AI in those moments. The question is whether it should be the first place you go.
For Josh Cote, it usually should not be.
AI Works Best Once You Have Something to Test
AI is particularly useful when there is already a point of view on the table.
You may have thought through a problem, spoken to somebody who has relevant experience and come to a provisional conclusion. At that stage, AI can be helpful for challenging assumptions, organising the argument, identifying gaps or helping turn a rough decision into something that can be communicated more clearly.
Used in that way, it becomes part of the thinking process without replacing the thinking itself.
The problem starts when somebody reaches for it before they have spent any real time with the issue. The answer may arrive quickly, but speed can disguise the fact that two valuable parts of the process have disappeared: the effort involved in forming your own judgment and the conversations that strengthen your professional relationships.
You Still Need to Learn How to Work Through a Problem
Good judgment rarely develops because somebody has immediate access to the right answer.
It normally develops through repetition. A person is faced with something they do not fully understand, considers the options, makes a call and then sees what happens. Sometimes they get it right. Sometimes they do not. Either way, they gradually become better at recognising patterns and making decisions when the information is incomplete.
That is a large part of what experience actually is.
When every difficult question is passed straight to AI, there is a risk that the uncomfortable part of that process gets removed. Somebody can still end up with a perfectly sensible answer, but they have not necessarily gone through the reasoning that would help them recognise the same kind of problem next time.
This is one of the concerns Josh has when founders become too dependent on AI at the beginning of a decision rather than later in the process. The work can become more polished without the person necessarily becoming better at making the judgment behind it.
There is a meaningful difference between using a tool to strengthen a conclusion and using it to avoid forming one.
The Other Cost Is Less Obvious: Fewer Conversations
Before AI tools became part of everyday business life, a difficult problem often meant speaking to somebody.
A founder might call another founder, a former colleague, an adviser or somebody who had already dealt with the same sort of situation. The immediate purpose of the conversation was to solve a problem, but something else was happening at the same time. The relationship was being maintained.
Those calls matter more than they sometimes appear to.
When you ask somebody for their view, you are telling them that their experience is useful to you. When they ask you for help later, there is already some history between you. Over several years, dozens of those small exchanges can become a strong professional network.
If every question that used to produce one of those conversations is now dealt with privately through a chatbot, there is a possibility that the network gradually becomes weaker without anybody consciously deciding to neglect it.
That matters because difficult business situations are rarely solved by information alone. Sometimes you need context. Sometimes you need somebody to challenge the premise of the question entirely. Sometimes what you really need is an introduction, a recommendation or a person willing to help when circumstances become messy.
AI can provide useful information, but it cannot replace the accumulated value of relationships built over time.
Research cited by the U.S. Small Business Administration has previously highlighted the value business owners place on mentoring and experienced support. The broader point is straightforward: people who have already lived through similar situations can often add something that generalised advice cannot.
That does not make AI less useful. It simply means it should not quietly replace the conversations that helped people build strong networks in the first place.
The Order Matters More Than the Tool
Josh’s argument is not that founders should stop using AI.
It is that there is a better sequence.
Start by thinking through the issue yourself. Work out what you currently believe, even if the view is incomplete. Then speak to somebody whose experience is genuinely relevant. Listen to where they agree, where they push back and what they notice that you may have missed.
Only after that does AI become particularly useful.
At that point, you can ask it to stress-test the reasoning, organise different options, identify missing questions or help you turn a decision into a clearer plan.
The technology is still doing valuable work, but it is operating alongside human judgment rather than standing in for it.
That difference is important. One process leaves you with an answer. The other can leave you with an answer, better judgment and a stronger relationship with somebody in your network.
How This Affects Conversations at EAW Consulting
At EAW Consulting, Josh increasingly encourages clients to arrive at strategy discussions with some initial thinking of their own.
That does not mean presenting a fully formed answer. In fact, a rough view can be far more useful than something that has already been polished into a finished-looking AI response. What matters is being able to explain what you currently think, what information shaped that view and whether you have discussed it with anybody who has relevant experience.
That gives the conversation somewhere meaningful to begin.
A founder who has already spent time wrestling with the issue is usually able to explain where the uncertainty really sits. They know which assumptions feel strong and which ones they are less sure about. The discussion can then focus on the parts that genuinely need challenging.
By contrast, an AI-generated response can sometimes look more complete than the thinking behind it actually is. The language may be confident and neatly structured, but that does not necessarily mean the underlying assumptions have been tested by somebody who understands the business or the people involved.
In consulting work, that difference becomes obvious very quickly.
The point is not to arrive with a perfect answer. It is to arrive having done enough thinking that the conversation can move the decision forward rather than simply reconstructing the reasoning from scratch.
A Useful Question Before Opening AI
One simple habit can help keep the balance.
Before asking AI for an answer to a difficult business question, ask yourself what you currently think.
You do not need a polished response. You just need enough of a view to explain why you are leaning in a particular direction.
If you cannot answer that at all, it may be worth spending another few minutes with the problem before asking a tool to solve it.
A second question is equally useful: who do I know who has already dealt with something similar?
Not everybody in your network will have the answer, and there is no benefit in calling people simply for the sake of it. But if there is somebody who has faced the same hiring issue, worked through a similar client problem or made a comparable strategic decision, that conversation may be worth having before asking AI to fill in the gaps.
The aim is not to make every decision slower. It is to make sure convenience is not gradually replacing the habits that build better decision-makers.
Why Founders Need to Be Particularly Careful
Founders are especially vulnerable to overusing AI as a first response because the role can be isolating.
There is often no manager sitting nearby to ask. There may not be another person in the business who understands the full context of the decision. Some problems involve information that employees cannot easily help with, while others are simply decisions the founder is expected to make alone.
AI fits neatly into that gap. It is available immediately and can produce something that feels like a second opinion in seconds.
The danger is allowing it to become a substitute for building the network that a founder is eventually going to need.
Strong professional networks are rarely created during a crisis. They are usually built gradually through conversations, introductions, favours, shared experience and people asking each other for advice long before anything urgent happens.
That is why Josh encourages founders to keep making those calls.
The person you speak to today about a pricing issue may later introduce you to a potential hire. Somebody you ask for advice on a difficult client may call you six months later with an opportunity. Those connections build in ways that are difficult to predict at the time.
A private conversation with AI does not have the same cumulative effect.
Use AI to Improve Thinking, Not Avoid It
There is no reason to treat AI as an enemy of good decision-making. Used properly, it can make decisions clearer, help expose weak reasoning and save time on work that does not require another person’s involvement.
The more useful distinction is between assistance and substitution.
If AI is helping you sharpen something you have already thought about, it is doing what the technology does well. If it becomes the place where every difficult question starts, there is a risk that the habits underneath good leadership begin to weaken.
For Josh, the preferred order is simple enough: think about the problem, speak to the right person if one exists, and then use AI to check or strengthen the conclusion.
The value of that sequence is not just the decision that comes out at the end. It is everything that gets built while reaching it: better judgment, more experience and stronger relationships.
Those are much harder to automate.
Josh Cote is the founder of EAW Consulting, working with business owners and leadership teams across the UK and US on strategy, growth and the habits that help leaders make stronger decisions over time.
