The short answer.
Plan how your company uses AI because a plan turns AI from something each person tries into something the company does on purpose. The week after a one-page plan exists, five things are different: you can say who uses AI for what, by team and by role; you know what gets built first and why; a good find in one team reaches every team that does the same work; you know which seats and tools are earning their place; and people hand over whole pieces of their own work instead of stopping at an assistant.
The first plan is one page, written in one sitting with the leaders of each team.
Without a plan, AI use is undirected.
Without a plan, a company’s AI use is undirected: the tools are switched on and paid for, and each person is left to work out what to do with them. Experimenting was the right way to start. Experimenting stops paying once it is all there is, and leaving AI to each person is still a decision, with a cost. What happens when a company uses AI without a plan? sets out that cost. The rest of this piece is about what changes once the plan exists.
You can say who uses AI for what.
With an AI plan, you can say who uses AI for what, by team and by role. Before the plan, the honest answer to “who uses AI here?” is a short list of enthusiasts. After the plan, the answer is a page: the finance team uses a workflow to prepare month-end notes, the account managers use one to prepare renewal briefs, and each workflow has one owner, by name, and one person who decides on its output. When someone new joins a role, the plan tells them which workflows come with the seat.
You know what gets built first, and why.
An AI plan puts the work in order: one workflow per team to start, and no more than two or three across the company at once. Each choice rests on what you can see today: how often the work happens, whether it starts from records someone can export, whether one person already reviews it, and what a miss would cost. When a team leader asks why their request waits, the plan gives the reason. When the first workflows are in daily use, the plan already says what comes next.
A good find reaches every team that does the same work.
With an AI plan, a good use of AI found in one team travels to every team that does the same work. Without a plan, the coordinator who works out a reliable way to prepare call notes keeps it in a personal chat history. On the plan, that find becomes a workflow with a name and an owner, listed in every team that prepares call notes. The work that crosses between teams goes on the page too, so a workflow for one side of a handoff is planned with the other side in the room.
You know which seats and tools earn their place.
An AI plan tells you which seats and tools are earning their place, because each one now points at named work. A seat held by someone whose role has workflows on the plan is earning its place. A seat with no planned work behind it is a question to ask. A tool that repeats what your main assistant already does is one to stop paying for. Every purchase after the plan has a place to land, or a reason not to happen. How do we get more value from the AI we already pay for? goes further on the seats you already hold.
People stop at “assistant” less often.
With an AI plan, people hand over whole pieces of their own work instead of using AI only to ask, draft and summarize. Left alone, almost everyone stops at the assistant, because nobody has shown them which parts of their job can go further. The plan shows each role which pieces of its own repeat work to hand over: the weekly status report, the first pass at a quote, the intake summary. A workflow prepares that piece the same way every time, in the ChatGPT, Claude, Copilot or Gemini you already pay for. The person checks it and decides. The larger gain sits there, and it happens only when someone plans it.
AI is now a real line in the budget.
After years of experimenting, AI is a real line in the budget, and it deserves the discipline and focus any spend of that size gets. You would not buy a new system for every desk and leave each person to decide what it is for. A plan gives AI the treatment any other investment gets: a direction, an order, an owner for each piece, and a way to see what it returns. Directed at named work, on purpose, the same seats do far more. That is how you 10X the ROI of your existing AI.
The first plan is one page and one sitting.
The first AI plan fits on one page and takes one sitting with the leaders of each team. Write one sentence on where the company is going with AI, list the teams and their roles, and give each team its first workflow with an owner by name. Decide where your company is going with AI shows how to write the sentence. The page will change as the work does. A short plan that stays current does more than a long one nobody opens.
Before the plan, after the plan: a sheet to fill in.
Fill in both lists for your own company. The first describes today. The second is what your one-page plan will say.
Before the plan
- Who uses AI here.________
- What we are building first, and why.________
- The best use of AI anyone has found, and who else knows about it.________
- What our AI seats and tools are for.________
- What most people use AI for.________
After the plan
- Each team, its roles and their workflows, each with an owner by name.________
- The first workflow for each team, and why it comes first.________
- Finds that became workflows, and every team they now reach.________
- Seats and tools that earn their place, and the ones we stop paying for.________
- The piece of repeat work each role hands over first.________
One firm writes its first plan.
Illustrative example.
Take a 70-person engineering services firm with an AI assistant on every desk. Before its plan, the leadership could name three heavy users and little else. The managing partner sat down for one afternoon with the leaders of project delivery, business development, finance and the office. They wrote one sentence about where the firm is going with AI, listed the four teams and their roles, and chose one workflow each for three teams: a weekly project status summary for delivery, first drafts of proposal sections for business development, and month-end billing notes for finance. The office team’s first workflow waits for the second round. Each workflow has an owner by name. One project manager’s own way of preparing status summaries became the starting point for the delivery workflow, and the same week the operations lead checked the seat list against the plan.
What is different in a company of under 200 people.
In a company of under 200 people, the people who decide fit in one room, so the first AI plan takes one sitting, not a programme. Most published advice on AI planning assumes a large enterprise: a steering committee, a pilot division, a year of review. You need none of it. The leaders of each team know the work, one person owns the plan (usually the owner, the head of operations or the chief of staff), and every team can have its first workflow in the same quarter.
Where the plan lives.
A first plan can live in a document. FullyPowered Planner is where it lives once your teams add to it: you set out your teams, their roles and the work between them, and choose what to build first, and your colleagues add their own teams. How organizations plan their use of AI walks through all three steps: plan it, develop the workflows, deploy them.