The Setup Last month, a client called me in a panic. They had six months of revenue data, three competing product lines, and no clear direction on which one to push. Their market was shifting. Mortgage rates were falling, which meant their customer base was about to change. They needed a strategy by end of week. Instead of running a three-day workshop, I decided to test something. I sat down with an AI tool, fed it their financials, customer feedback, and market context, and asked it to generate a full 10-section business strategy. Start to finish. Timed it. 28 minutes. A complete first draft. Here is what happened next, and why the speed matters far less than what you do with it. What the AI Actually Did The tool generated sections on market positioning, revenue priorities, operational bottlenecks, competitive differentiation, cash flow projections, resource allocation, customer segmentation, pricing strategy, risk assessment, and a 12-month roadmap. Was it perfect? No. Was it complete enough to work from? Absolutely. This is the gap most people misunderstand about AI in strategy work. The tool did not think. It synthesised. It took patterns from thousands of business plans, matched them to the inputs I gave it, and assembled a structure that was logically sound. It was like having a very fast junior consultant who had read every business textbook but had never sat in a room with a real business owner. That is actually useful. But only if you know what happens next. The Real Work Begins My client and I spent the next four hours doing what the AI could not do. We read through every section and asked hard questions. The strategy said revenue should shift 60% toward Product B. I asked why. The AI had based this on market size data and margin analysis. Reasonable. But my client knew that Product B had a three-month sales cycle and their cash position could not absorb that transition. The AI did not know that. We changed the recommendation to 40% shift, phased over nine months. The competitive positioning section was generic. It said they had "strong customer service and flexible pricing". Every business says that. We rewrote it to focus on something true and specific. They had zero customer churn in one segment because they had built the product around that segment's exact workflow. That was the real differentiation. The AI had missed it because it was not in the financial data. The risk section flagged supply chain disruption and market volatility. Fair points. But it did not flag the real risk, which was that their operations manager was planning to leave in three months and they had not trained anyone to replace her. That risk would kill the strategy dead. The AI could not see it because I had not mentioned it in the brief. Why This Matters Right Now Markets are unstable. Oil prices swing on geopolitical news. Interest rates move on bank announcements. If you are running something small, you cannot afford to spend three weeks on strategy work. You need a rough map fast, then you need to sense-check it against what you actually know. That is where AI works. Not as a replacement for thinking. As a way to compress the busywork so you can focus on the thinking. The old way: hire a consultant, spend two weeks in workshops, get a 40-page strategy document you half-read. Cost. Time. Overhead. The new way: use AI to generate a first draft in half an hour, spend four hours tearing it apart and fixing it, move forward with something real. You save roughly 80 hours of workshop time and consultant fees. That matters when you are bootstrapped or running lean. Where AI Falls Short I need to be straight about this because I have built AI tools and I know where the gaps are. AI cannot see tacit knowledge. It cannot know that your best customer is about to leave, or that your team is burnt out, or that a competitor is hiring aggressively in your territory. It cannot understand what happens in the room when people talk. It cannot feel the market. AI cannot make the hard trade-offs. It can tell you that you should pursue two different markets simultaneously. A real strategy requires you to say no to one of them so you can be excellent at the other. That choice belongs to you. AI cannot account for your constraints. It does not know your personal risk tolerance, your financial position, your network, or what you actually want to build. A strategy that looks good on paper but asks you to spend money you do not have or take risks you will not take is worthless. AI also hallucinates specifics. It will cite market research that does not exist or give you numbers that sound real but are invented. You have to fact-check everything. What Actually Happened With My Client They took the AI draft, worked through it with me, and ended up with a strategy that was 40% AI-generated framework and 60% their own thinking. They implemented the phased shift toward Product B. They rewrote their positioning. They got clear on their resource constraints and built a roadmap that actually fit their cash position. Six months later, they had grown revenue by 23% and their team was still intact. They did not burn out trying to do three things at once. They did the one thing that mattered. Would they have got there without the AI draft? Maybe. But they would have spent longer debating the basics instead of focusing on the decisions that actually changed the outcome. What to do this week If you need a strategy or a plan and you have been putting it off because it feels like a three-week project, try this. First, spend 30 minutes writing down your business context. Revenue by product line. Top three customer segments. Your biggest operational bottleneck. One market shift you are watching. Your cash position for the next 12 months. Feed this into an AI tool with a prompt asking for a 10-section business strategy. Second, print the output and read it with a pen. Mark every section that is generic, wrong, or incomplete. Mark the sections that actually land. You should have questions. Third, book four hours with someone who knows your business but is not inside it every day. Work through the marked-up strategy together. Fix it. Add the stuff the AI missed. This is where it becomes real. Do not wait for perfect. Do it this week.