The honest version I built AI-powered business tools. I use them with clients. And I will tell you directly: they are not as good as people want them to be. That is not a criticism of the technology. It is a statement about what the technology actually does versus what people expect it to do. The gap between the demo and the reality is where most of the confusion lives. AI strategy tools are genuinely useful. But they are useful in specific ways, for specific problems. The moment you treat them as a replacement for your own thinking, you have made a mistake. What AI tools actually do well They impose structure. That is the core value. When someone comes to me with a decision they cannot make, the problem is rarely that they lack information. The problem is that the information is scattered across their head, their inbox, three conversations they had last week, and a spreadsheet they started but never finished. A tool like the SWOT Analysis at alira.london does not tell you what your strengths are. It asks you to write them down in a format that forces comparison. The 5 Whys tool does not diagnose your root cause. It makes you keep asking until you stop giving surface-level answers. This is valuable. I have seen people solve problems in 20 minutes that they had been circling for months, simply because they were forced to articulate the issue in a structured way. But the tool did not solve the problem. The structure did. The person did the actual thinking. Where the tools fall short AI cannot tell you what matters to you. It cannot weigh your risk tolerance, your cash position, your relationship with a specific supplier, or whether you actually want to grow or just want to stop feeling overwhelmed. I worked with someone recently who ran a Decision Matrix on whether to take on a major contract. The matrix said yes. Every factor they entered pointed toward accepting. But when I asked them how they felt about the result, they said they felt sick. That reaction was information. The tool had no access to it. The contract would have required working with someone they did not trust. They had not entered that as a factor because they felt embarrassed about it. It seemed unprofessional to reject a contract based on a gut feeling about a person. They turned it down. Six months later, that person's business collapsed owing money to half their suppliers. The matrix was not wrong. It processed what it was given. But it could not ask the question that mattered: what are you not telling me? The 70% rule Here is how I think about it: AI tools can get you 70% of the way to a decision. They can organise your thinking, surface factors you had not explicitly named, and give you a framework to compare options. The last 30% is yours. That is where context lives. Where relationships live. Where the thing you know but cannot prove lives. If you try to outsource that 30%, you will make decisions that look rational on paper but feel wrong in practice. And decisions that feel wrong tend to get executed badly, even when they are technically correct. When to use the tool Use it when you are stuck in loops. When you have had the same conversation with yourself four times and gotten nowhere. When you cannot see the shape of the problem because you are too close to it. Use it when you need to communicate a decision to someone else. A structured analysis is easier to explain than "I just know." Even if the knowing came first and the structure came after. Use it when you are comparing genuinely similar options. Two suppliers with similar pricing, two marketing approaches with similar costs, two hires with similar experience. The tool helps you name the small differences that actually matter. Use it when you are prone to overthinking. Some people will research forever if you let them. A framework that says "here are the five factors, now score them" creates a forcing function. It makes you decide. When to trust yourself instead Trust yourself when your body is giving you a signal your brain is trying to override. Nausea, dread, relief. These are data points. They are not always right, but they are always worth investigating. Trust yourself when you have direct experience the tool cannot capture. You have worked with this client before. You know how they communicate. You know what they actually mean when they say "flexible timeline." No framework can hold that knowledge. Trust yourself when the stakes are existential. The political uncertainty this week, with borrowing costs rising and the pound falling, is a reminder that external conditions can shift fast. If a decision could sink you, do not let a matrix make it. Use the tool to think. Then sit with the answer before you commit. Trust yourself when everyone else is doing the same analysis. If your competitors have access to the same tools and the same data, the tool-generated answer is the average answer. Differentiation comes from the part only you can see. The actual skill The skill is not using AI tools. Anyone can fill in a template. The skill is knowing when the template output is the answer and when it is just a starting point for the real conversation. I use the tools at alira.london with clients regularly. But I have never once said "the tool says X, so do X." I say "the tool says X, what do you think about that?" The answer to that question is where the value lives. What to do this week Pick one decision you have been putting off. Run it through a structured framework. The Decision Matrix at alira.london takes about 15 minutes. Write down the result. Then close the tab and ask yourself: does this feel right? If yes, act. If no, write down what the tool missed. That gap is your 30%. Finally, notice your pattern. Do you tend to over-rely on analysis, or do you tend to skip it entirely? Whichever direction you lean, try the opposite this week. Just once. See what happens.