The short answer.

An AI strategy says why your company uses AI and where it matters most. An AI policy says what is allowed: which assistant, what information may go into it, and that a person decides. An AI plan says which team and which role gets which workflow, in what order, owned by whom by name. A company of under 200 people needs a strategy of a few sentences, a policy of two pages, and a plan it keeps current, because the plan is the only one of the three a team lead can act on next Monday.

The three documents answer three different questions, and most companies have written one and believe they have all three. A company with only a policy has permission without direction. A company with only a strategy has direction without anyone assigned to it. The plan is where the strategy turns into work, and it starts the three steps that follow: plan it, develop the workflows, deploy them.

An AI strategy says why and where.

An AI strategy states why the company uses AI and which parts of the business it should help first. It is a choice of direction, made by the owner or the leadership team, and in a company of under 200 people it fits in a paragraph.

  • Why.The reason in the company’s own terms: more time for clients, faster answers to customers, fewer evenings spent on reports.
  • Where.The two or three parts of the business that matter most for that reason, named by team or by kind of work.
  • What it leaves out.The work the company keeps with people on purpose, said once.

An AI strategy names no assistant, no workflow and no owner.

An AI policy says what is allowed.

An AI policy sets the rules everyone follows whatever work they do with AI. It names the approved assistant, such as ChatGPT, Claude, Copilot or Gemini on the company’s own account; says which information may and may not go into it; says that a person checks and decides before anything is sent, paid or changed; and names who to ask. Does a small company need an AI policy? covers what goes in those two pages.

An AI policy is the same for every team. It assigns no work to anyone. A company that writes its policy and stops has made AI safe to use and left every person to decide alone what to use it for. This page is general guidance, not legal advice; take advice on your own policy where your sector or your data calls for it.

An AI plan says who runs what, in what order.

An AI plan lists the company’s teams, the roles in each team, and the workflows each role will run, in the order they will be built, with one named owner for each workflow and a date to review. The AI plan is the document that turns the strategy’s “where” into named work, inside the policy’s limits.

  1. Teams first.Each team, named for the work it does, with its lead.
  2. Then roles.The roles in each team, because a workflow belongs to a role and stays when a person moves. How to plan AI by role goes deeper.
  3. The workflows.The pieces of work each role repeats, and which one is built first.
  4. The order.Which team starts, which comes second, and why.
  5. The owners.One person by name for each workflow, who keeps it current and decides when it changes.
  6. The review date.When the plan is read again and changed.

The AI plan changes every month as workflows are built and new ones are chosen. The strategy and the policy change once or twice a year. How to write an AI plan in one sitting gives the meeting that produces the first version.

How to tell which document you are holding.

Ask what the document lets a team lead do on Monday morning. The answer tells you which of the three you have.

  • It tells the lead what matters.That is a strategy.
  • It tells the lead what they may and may not do.That is a policy.
  • It tells the lead which workflow their team builds next and who owns it.That is a plan.

A document titled “AI plan” that lists tools to buy and names no team, role, workflow or owner is not a plan.

All three on one page, to copy.

Copy these headings onto one page. Fill in the strategy and the policy once. Fill in the plan section for each team and keep it current.

  1. Strategy: why we use AI.One or two sentences.
  2. Strategy: where it helps first.Two or three teams or kinds of work.
  3. Strategy: what stays with people.The work we keep with people on purpose.
  4. Policy: the approved assistant.Which one, on which account.
  5. Policy: information.What may go in, and what never does.
  6. Policy: who decides.A person checks before anything is sent, paid or changed.
  7. Policy: who to ask.One name.
  8. Plan, per team: team and lead.The team’s name and its lead.
  9. Plan, per team: roles.The roles in the team.
  10. Plan, per team: workflows in order.First, second, third, with the role that runs each.
  11. Plan, per team: owner.One name for each workflow.
  12. Plan: review date.The day the plan is read again.

One school writes all three.

Take a 70-person private school. Teaching staff are most of the headcount; the office side is an admissions team, a business office and a front office. The head of school wrote all three documents with the director of admissions and the business manager.

Illustrative example.

  • The strategy.AI helps the office give families quick, personal answers and gives staff more time with families in person. Admissions and the business office come first. Anything said to a family about a child’s admission decision or wellbeing stays with people.
  • The policy.The school’s approved assistant, on the school’s own account. No student records, health information or family financial details go into any personal account. A named member of staff reads every message before it reaches a family. Questions go to the business manager.
  • The plan, admissions team.Roles: director of admissions and admissions coordinator. First workflow: drafting replies to enquiry emails from the school’s own approved answers, run by the coordinator and owned by the director. Second: a checklist of what is missing from each application file, prepared from the admissions system’s export.
  • The plan, business office.Roles: business manager and accounts assistant. First workflow: drafting tuition statement follow-ups for the business manager to approve, owned by the business manager. It waits until the admissions workflow is in daily use.
  • The plan, front office.Later, once the first two are running, with the weekly family newsletter as the candidate.

The plan, not the strategy or the policy, told the admissions coordinator which workflow was hers and who owned it. The enquiry reply workflow starts from a free workflow template in the Workflows Directory, adapted to the school’s own approved answers.

What is different in a company of under 200 people.

In a company of under 200 people, the strategy and the policy are short and the plan carries the weight. Most published advice is written for large enterprises, where strategy and policy run to many pages. You do not need that.

  • The strategy fits in a paragraph.The owner already knows why and where. Write it down so every lead reads the same words.
  • The policy fits on two pages.One approved assistant, the information rules, a person deciding, one name to ask.
  • The plan is the document you return to.Every lead can see their team, their roles and their next workflow, and the owner of each workflow is someone everyone knows by name.

After years of experimenting, that plan is what makes AI use directed instead of undirected, and it is how you 10X the ROI of your existing AI.

Write the strategy and the policy this week, then put the plan where everyone can see it. FullyPowered Planner is where the plan lives: you set out your teams, their roles and the work between them, choose what to build first, and your colleagues add their own teams. How organizations plan their use of AI covers the rest of the method.