What should you actually do about AI?

A founder at a networking event said she needed to "get an AI strategy." When asked what problem she was trying to solve, she paused. "I'm not sure yet. But everyone says we need to be using it."

That's the whole problem in two sentences.

AI is a tool, not a strategy

A strategy has a goal. "Use AI" is not a goal — it's a category of tools, like "use software." The question worth asking is narrower: what is slow, expensive, or unreliable in your business right now, and would AI do it better than what you're currently doing?

Start there. The answers are usually less exciting than the pitch decks suggest, and considerably more useful.

Where it genuinely helps

A few places where AI consistently delivers for small businesses, without requiring a transformation programme:

First drafts. Proposals, emails, social posts, job descriptions. AI doesn't replace the judgment about what to say — it removes the blank-page problem. Most founders spend far too long on documents that are structurally similar to ones they've written a dozen times before. AI handles the structure; you handle the substance.

Summarising long content. Contracts, reports, transcripts of meetings you missed. Reading a 40-page document to find the three things that matter is expensive use of a senior person's time. It doesn't have to be.

Repetitive customer queries. If the same ten questions come in every week, an AI-assisted response layer handles them at low cost. Not appropriate for sensitive conversations — but for "what are your rates?" and "how do I cancel?", it works.

Research and first-pass analysis. Market overviews, competitor summaries, briefing documents before a meeting. What comes back isn't the finished work — but it turns an hour into fifteen minutes.

Where it doesn't

AI produces plausible-sounding output. That's not the same as correct output. Anywhere the cost of being wrong is high — legal documents, financial projections, anything you're putting your name to — AI is a starting point that requires a human to verify, not a finished product.

It also doesn't replace relationships. A client who emails you wants to know a person read it. Sales conversations, complaint handling, difficult negotiations — the value is human judgment and accountability. Automating those is a false economy.

The practical starting point

Pick one thing that takes too long, happens often, and doesn't require much judgment to get right. Try an AI tool on that task for a month. Measure whether it actually saves time and whether the output is good enough. If yes, expand. If no, move on.

That's it. That's the AI strategy.

The businesses getting real value from AI aren't the ones with transformation roadmaps. They're the ones who found a specific, tedious task and stopped doing it by hand.

If you want help identifying where AI can actually move the needle in your business, book a conversation.