The short answer.
Not having an AI plan is still a plan. A company without one runs on a default plan that says each person works AI out alone, and that default makes real decisions for you: which work moves onto AI, who owns it, what good looks like and what is allowed. You replace the default by writing down what it decides today, then making each of those decisions yourself, on one page, with one owner.
Experimenting was the right way to start, and nobody chose the default on purpose. It arrived with the seats. What happens when a company uses AI without a plan? sets out the visible result; this piece names the decisions behind it.
What the default plan decides for you.
The default AI plan makes six decisions every month, whether or not anyone signs them. Each one has an answer today; the default just never wrote it down.
- Which work moves onto AI.The default says: whatever each person happens to try. That is usually drafting and summarizing, and almost never the work a team repeats every week.
- Who owns a method that works.The default says: the person who found it, and nobody else. When that person is on leave or moves on, the method goes with them.
- What good output looks like.The default says: whatever each person accepts. Two people doing the same task produce two standards.
- Who checks the work.The default says: the person who made it, if they remember. Nobody designed the check.
- What is allowed.The default says: each person guesses. Some paste company records into a personal account; others avoid AI entirely in case they are doing something wrong.
- What the seats are for.The default says: nothing in particular. The renewal arrives, and the only answer anyone can give is that some people use it for some things.
Each is a decision nobody was asked to make.
What the default feels like from the inside.
Living under the default AI plan feels confusing for almost everyone in the company, and each role carries a different part of that confusion.
- For the people doing the work.The first weeks are exciting. Then the questions pile up with nowhere to take them: am I using this right, am I allowed to put this in, is everyone else further ahead than me? Some keep going quietly. Some stop and feel they fell behind.
- For the team leads.A lead sees one person racing ahead and another not touching it, and has no basis to say what the team should do. Every answer feels like a personal opinion.
- For the owner or leader.You approved the spend, you hear it mentioned in meetings, and you cannot say what it is doing for the business. That unease is reasonable. It is the default plan reporting back.
The experience is chaotic; the people in it are doing their best with no direction. That is the cost the default never shows on an invoice.
Where the default plan gets complicated.
The default AI plan gets complicated wherever work passes from one person to another, because every person in the chain runs a different private method.
- At the handoff.One person prepares work with AI their way; the next person cannot tell what the assistant was given or what it left out, so they check everything again.
- In front of customers.The same kind of document goes out in five shapes, depending on who wrote it.
- For new hires.A new person asks how the team uses AI and gets a different answer from each colleague.
- When a good find leaves.The best method in the company sits in one chat history, and the company finds out it mattered when it is gone.
Every new hire adds one more private method.
Replace the default one decision at a time.
You replace the default AI plan by taking its six decisions back, one at a time, and writing your answer down where everyone can see it. You stop nothing. People keep experimenting, now inside a plan that says where the company is going.
The replacement is one page. It names your teams and the lead of each, the roles in each team, the work each team repeats, the one piece of work per team that moves onto AI first, and one person, by name, who owns each workflow and the page itself. How to write your company’s AI plan in one sitting runs the meeting that produces it.
FullyPowered Planner is where that plan lives. You set out your teams, their roles and the work between them, choose what gets built first and name who owns it, and colleagues add their own teams.
Each first piece of work then becomes a workflow: your best way of doing that work, written down precisely enough that an AI assistant prepares it the same way every time, tested on real cases, and run in your company’s own ChatGPT, Claude, Copilot or Gemini by your own people. A person decides; the workflow prepares. Where the method sits in one expert’s head, FullyPowered Studio captures it as a workflow.
The default plan, written down: a sheet to copy.
Copy these six rows onto one page. In the first column, write what happens today, honestly. In the second, write your decision and the name of the person who holds it.
- Which work moves onto AI first.Today: ______. Our decision: one piece of repeat work per team, named.
- Who owns each method that works.Today: ______. Our decision: one owner per workflow, by name.
- What good output looks like.Today: ______. Our decision: written in the workflow, one standard per piece of work.
- Who checks the work before it goes out.Today: ______. Our decision: one named reviewer per workflow.
- What is allowed.Today: ______. Our decision: company accounts only, and where to ask when unsure.
- What the seats are for.Today: ______. Our decision: the workflows each team runs, listed by team.
At the top of the page, write the name of the person who owns it and the date you will look at it again.
A staffing agency replaces its default.
Take a 110-person staffing agency with four teams: recruiting, client sales, onboarding and compliance, and payroll and billing.
Illustrative example.
Under the default, every recruiter wrote candidate summaries for clients with their own prompts, so clients received summaries in a dozen shapes. One onboarding coordinator had worked out a careful method for chasing missing documents; nobody else knew it existed. Some people were unsure whether candidate details could go into an assistant at all, and stopped using it. The managing director felt the confusion in every leadership meeting.
The managing director and the head of operations filled in the six-row sheet with the four team leads. Recruiting chose the candidate summary as its first workflow, owned by the head of recruiting, with each recruiter approving their summary before it goes to a client. Onboarding chose the missing-documents follow-up, and the coordinator who found the method wrote it down for the team and owns it. Payroll and billing chose the weekly timesheet exceptions review, owned by the payroll lead. Client sales went on the page as the team to look at next. Everyone moved onto the company account, and the head of operations became the owner of the page, reviewed on the first Monday of each month.
What is different in a company of under 200 people.
In a company of under 200 people, the default AI plan is easier to see and quicker to replace. Most published advice is written for large enterprises, where an AI plan means a steering group and a long program.
- You know whose method is whose.The good finds have names on them, and you can ask each person directly to write theirs down.
- The plan fits on one page.Every team lead can sit at one table and settle the six decisions together.
What do we tell employees about AI? covers how to announce it so people feel the relief of direction, not a rebuke.
Write down your default, then replace it. How organizations plan their use of AI sets out the three steps: plan it, develop the workflows, deploy them. FullyPowered Planner is where your plan lives once it is written.