A One-Page AI Plan Any Owner Can Write in an Afternoon

A simple template that turns “we should use AI” into a concrete, sequenced plan.

Below, we keep it practical and free of jargon: what is actually going on, how to think about it for a small or medium-sized business, the steps to take, what to watch out for, and an honest read on what to expect. The throughline is the same one we bring to every engagement — use AI to help your team get more done and work at a higher level, and make sure your people genuinely know how to use the tools.

The state of play

AI stopped being a future-tense topic for small businesses a while ago. According to research, 88% of organizations now report using AI in at least one business function (McKinsey, “The State of AI,” 2025). The real gains come from applying a few of these tools well to the work that slows your team down, then teaching your people to use them confidently. This piece is about doing that deliberately, rather than chasing whatever is trending this week.

A practical framework

The owners who get real value from AI share a habit: they pick a single, painful, repetitive task and fix it completely before touching anything else. They keep a human reviewing the output, they make sure the team actually knows how to use the tool, and they measure one number. Once the first win is undeniable, they expand. It is unglamorous, and it works.

1. Name the task. Choose one repetitive, high-frequency job that drains time or invites errors.
2. Map the steps. Write down how it is done today, including the judgment calls a person makes.
3. Insert AI at one step. Let it draft, sort, extract, or summarize, with a human still approving the result.
4. Teach the team. Walk the people who do this work through using the tool, so it sticks.
5. Measure for two weeks. Track time saved or errors avoided against your one number.
6. Expand or stop. If it is working, widen the scope. If not, you have lost two weeks, not two quarters.

Why it pays off

The point is not to shave a few minutes off a task. It is to change what your team can take on. When the repetitive layer is handled and people are comfortable with the tools, the same headcount starts producing more, turnarounds get faster, and attention moves to the work that grows the business. Research suggests the potential is real: Generative AI could add the equivalent of $2.6–$4.4 trillion in value annually across 63 use cases (McKinsey Global Institute, 2024). That is a description of what is possible across many organizations, not a promise about yours.

Where it goes wrong

• Buying a tool before you have defined the problem it is solving.
• Automating a broken process instead of fixing it first.
• Skipping the review period and trusting the output blindly.
• Spreading effort across ten experiments instead of finishing one.
• Rolling out a tool without training the people who have to use it.
• No owner assigned, so the automation quietly rots.

What to expect

It is worth saying plainly: AI is not magic, and adoption has a learning curve. Plenty of projects stall, almost always because no one owned them or the team was never brought along. Treat your first automation as an experiment with a clear owner, a clear metric, and time set aside to teach the people involved. Do that, and you avoid the most common way this goes sideways.

Common questions

Bottom line: Pick the most painful version of this problem, fix it first, and build momentum from a win your people can see.