AI Strategy for Small Businesses: A One-Page Plan

Most AI strategy documents are written for large companies with data teams and transformation budgets. A business with five or fifty people needs something different: a plan short enough that the owner actually reads it, specific enough to act on on Monday, and honest about what will be measured.

Below is the template I use with small business clients, section by section, followed by a filled-in example for a made-up business so you can see how it reads.

Why does a small business need an AI strategy at all?

A small business needs a short AI strategy because without one, AI use becomes a scatter of personal subscriptions and half-finished experiments that nobody measures. A plan focuses limited time and money on one problem worth solving.

You are not late. The US Census Bureau's business survey put the national share of firms using AI at 19.8% as of May 2026, with use among firms of four or fewer employees under 20% and not rising significantly, compared with 32% of firms with 100 to 249 staff. Individual use runs ahead of company plans: a Federal Reserve note estimates that about 41% of workers use generative AI for work. In many small businesses, staff already use AI on their own. A strategy turns that into something the business controls.

The one-page AI plan template

The template has six boxes. Each one has a single question to answer, and none should take more than a few lines.

How to fill in each box

Goal. Start from a business problem, not from AI. "Reply to new enquiries faster", "cut the time spent on monthly invoicing", "stop losing quotes in the inbox". If you cannot name the problem, AI is not the next step; a conversation with your team about where time goes is.

First use cases. Pick tasks that happen often, follow a pattern, and where a person can check the result easily. Drafting replies, summarising calls, pulling details from documents and sorting incoming requests are common first wins. Leave anything that moves money or makes promises to customers on its own for later. My guide to AI automation for small business goes deeper on choosing.

Data and tools. List what each use case needs to read: past emails, a price list, a help document, a spreadsheet. Note where each lives and who owns the account; who owns your automations explains why that matters. Sorting this out early saves the most time later.

Budget. Set two numbers: a monthly cap for subscriptions and usage, and any one-off cost for setup or a build. For scale, as of late 2026 Anthropic listed its Claude Team plan at $25 per standard seat per month on monthly billing, with a two-seat minimum. A tool budget for a small team can be modest; the larger cost is usually setup time, whether yours or a consultant's.

Risks and rules. Write down what could go wrong and the rule that prevents it. Typical rules: a person reviews anything sent to a customer, no personal or financial data pasted into personal AI accounts, and business accounts owned by the business rather than by one employee.

90-day measures. Pick three numbers and record today's value before you start: hours spent on the task per week, response time, error rate, or enquiries converted. Without a baseline, day 90 becomes a matter of opinion. My post on AI automation ROI explains how to turn those numbers into a payback figure.

A worked example (a made-up business)

Here is the template filled in for a hypothetical eight-person bookkeeping practice. It is an illustration, not a client, and the numbers are placeholders to show the format.

Notice how little technology appears on the page. The plan is about a business outcome, two tasks and how success will be judged. The tools follow from that, not the other way round.

What happens at day 90?

At day 90 you compare the three measures with their baselines and make one of three decisions: keep and extend, fix and retest, or stop. Then you rewrite the page with the next goal and use case, reusing whatever accounts, data clean-up and rules you already set up.

Three things are worth knowing before you get there:

• The second use case is cheaper than the first, because the accounts, data and rules already exist.
• Box 3 is where plans usually stall, because data turns out to be scattered or locked in one person's account. Fix that first.
• Unmeasured wins get abandoned. If nobody recorded the baseline, the tool gets cancelled at the next budget review, even when it helped.

If you want help with the first page, a short paid discovery is usually enough. I describe what one should produce in what an AI consultant delivers in 30 days, and typical prices are in how much an AI consultant costs.

Work with me

I help small businesses turn a one-page plan into working automations and AI features their team actually uses. See workflow automation and AI integration, browse my work, and book a call to work through your first page together.