The problem is never the technology. It is almost always the brief.
We speak to small business owners every week who want to "use AI" in their business. That instinct is right. But the first question they ask is almost always wrong: "What can AI do?"
That question leads to demo-driven thinking. Someone sees a chatbot, a document scanner, or an auto-generated report and decides they need one too. Three months and several thousand pounds later, they have a tool that nobody uses because it solves a problem nobody actually had.
Flip the question
The businesses that get real value from AI start from the other end. They ask: "What costs us the most time?" or "Where do we keep making the same mistakes?"
That is a completely different starting point. Instead of shopping for technology, you are mapping your actual pain. And once you have a clear pain point, the right solution often turns out to be simpler than you expected.
One of our clients was spending 12 hours a week manually matching purchase orders to invoices. They assumed they needed a full AI document processing system. What they actually needed was a straightforward automation that pulled data from two spreadsheets and flagged mismatches. No AI required. Problem solved in a week.
The three-line test
Before any conversation about AI, do this:
- Write down your three most time-consuming manual processes
- Estimate the hours per week each one takes
- Ask yourself: does a human genuinely need to do this, or are they doing it because nobody has built an alternative?
If you can answer those three questions honestly, you already have a better brief than 90% of the AI projects we see. The technology conversation comes after, not before.
Start with the problem. The solution will follow.