Practical guide
Does my small business need an AI strategy?
The short answer
Probably not, at least not in the sense of a long strategy document. Most SMEs get further with three things: a short list of business problems worth solving, simple rules for how staff may use AI tools, and one small, measured trial. The UK government's 2026 AI adoption survey found that 65% of businesses not using AI said they lacked a clear business case or proven return on investment. Starting from the problem fixes that.
Why the question comes up
Owners are told they are falling behind on AI. Suppliers offer strategy workshops. Staff are already using AI tools, sometimes without telling anyone. It is natural to feel you need a plan.
The numbers suggest adoption is real but not universal. The Office for National Statistics found that 30% of UK businesses were using at least one AI technology in September 2026, up from 9% three years earlier. The government's own AI adoption survey, which uses a different method, put it at 32%, with skills and the lack of a clear business case as the main barriers.
What you need instead of a strategy
1. A list of problems, not technologies
Write down where the business loses sales, time or money. For example:
- Quotes take too long and prospects go elsewhere.
- The same client information is chased every month.
- Data is typed into three different systems.
- The owner is the bottleneck for routine approvals.
Then ask which of those could be improved, and by what. Sometimes the answer is AI. Often it is ordinary automation, better use of existing software or a clearer process.
2. Simple rules for using AI tools
Your team is probably using AI already. A one-page policy is more useful than a strategy:
- Which tools are approved, and on which accounts. Business plans, not personal ones.
- What information must never be entered, such as client personal data or confidential documents, unless the tool is approved for it.
- That a person is responsible for checking anything AI produces before it is used.
- Who to ask when unsure.
Our guide on using ChatGPT safely with client data has a fuller checklist.
3. One measured trial
Pick one problem from your list, try one change for a month, and measure the result: hours saved, response time, errors avoided. Keep it, change it or stop it based on what happened.
When a more formal plan does make sense
A fuller AI plan is worth it when:
- AI will touch regulated decisions or large amounts of personal data.
- You are making significant investment across several teams or systems.
- Clients, insurers or regulators are asking how you use AI.
- You are building AI into the product or service you sell.
Even then, keep it practical: priorities, risks, responsibilities and a timetable, not a slide deck.
The cost of getting this wrong
The risk is not being late to AI. It is spending money on tools that change nothing, or letting sensitive information leak into services nobody approved. Both are avoided by starting with the problem.
A practical first step
The FiveW Business Review is designed as the alternative to an AI strategy. It looks at up to three important processes and gives you prioritised recommendations covering likely benefit, cost, effort and risk, including when AI is not the right answer. It costs £750 plus VAT and the report is yours to use with anyone.
The FiveW Business Review
Want this answered for your own business? The £750 + VAT review looks at up to three important processes and gives you prioritised recommendations covering likely benefit, cost, effort and risk.
Tell us what is happeningSources
Prices and figures were checked on 8 November 2026 and can change. Always confirm with the supplier.

