The short answer.

Your team is uneasy about AI because three questions are open for each person: am I allowed to use it for this, is my way the right way, and what does it mean for my role. Nobody has answered them, so each person carries them alone. A plan answers all three for every team, in writing, and the unease eases once people can read the answers.

Answering those questions is part of planning AI, the first of three steps: plan it, develop the workflows, deploy them. What happens when a company uses AI without a plan? describes the company’s view. This piece describes the view from your team’s desks.

What undirected AI feels like from your team’s side.

Undirected AI use feels to the people doing it like guessing in private. The company pays for ChatGPT, Claude, Copilot or Gemini, or people use their own, and nobody has said what it is for. Experimenting was the right way to start, and your people did it: they tried drafting, summarizing and checking, and found a few things that work.

What they did not get is any answer back. A billing coordinator who uses AI to draft a follow-up letter does not know whether that is welcome, tolerated or a problem. A colleague who never touches it does not know whether that is fine or a sign of falling behind. Both feel the same thing: the rules exist somewhere, and nobody has told them. The experience is confusing, and at times chaotic, for capable people doing their best.

What the unease costs the company.

Unease about AI costs the company the work people do quietly and the work they never try. When people are unsure, they behave in predictable ways.

  • Good finds stay hidden.A person who found a better way to prepare a weekly report keeps it to themselves in case it is not allowed. The team never gets it.
  • Work is done twice.People draft with AI, then redo the draft by hand because they are not sure the AI version will pass. The company pays for both.
  • Questions land on the lead.Every small doubt goes to the team lead, one conversation at a time, and each lead gives a different answer.
  • Careful people hold back.The people most careful about quality are the ones who wait longest, so the team’s best judgment is the last to reach its AI use.

None of this shows up on a report. It shows up as a team that seems slower and quieter about AI than you expected, and as a leader who cannot say what the company’s AI seats are for.

Am I allowed to use it for this?

A team knows what AI use is allowed when the plan says, for that team, which work AI prepares, which information stays out of it and which assistant the company uses. Three lines answer it. People stop guessing because the answer is on a page they can find, not in a lead’s head.

Keep the answer about the team’s own work. “Use AI responsibly” leaves the question open. “The scheduling team uses the company’s assistant to draft appointment reminders; patient records stay in the practice system” closes it. Does a small company need an AI policy? covers the company-wide rules; the plan carries the line for each team.

Is my way the right way?

A person knows their way is the right way when the team has one shared workflow for a piece of work and a named person who checks the output. Until then, every person runs the same work differently and nobody can tell whose version is good. A shared workflow replaces private guessing with one tested way, written down, that everyone on the team runs.

This is where the plan turns into something people use. FullyPowered Planner is where the plan lives: your teams, their roles, the work between them, what gets built first and who owns each piece. Each person can open it and see their team’s line. When a workflow is ready to build, we develop it with you, and FullyPowered Studio captures the way your best people already do the work.

What does it mean for my role?

A person knows what AI means for their role when the plan names, for that role, the work a workflow prepares, the decisions the person still makes and the workflow they own or check. Write it for the role, not for the company. People read “AI will help us all” as an evasion and “the billing coordinator checks every claim the workflow prepares before it goes out” as an answer.

A person decides; a workflow prepares. Said plainly for each role, that sentence tells people where their judgment sits. How to plan AI by role sets out how to write each role’s part of the plan.

Three answers to write for each team.

Copy these headings onto one page per team. Fill them in with the team’s lead, then read the page to the team and ask what is missing.

  1. What AI is for on this team.The pieces of work AI prepares here, by name.
  2. What stays out.The information that never goes into an assistant, and the system it stays in.
  3. Which assistant.The one the company uses, on the company’s account.
  4. The shared way.The team’s first workflow, or “none yet” with the date it is chosen.
  5. Who checks.One name for each piece of work AI prepares.
  6. Each role’s part.For every role on the team: what the workflow prepares, what the person decides, what the person owns.
  7. Where questions go.One person, by name.
  8. Where to bring a good find.One person, by name, and when the team reviews finds.

One dental group answers its three teams.

Take a 60-person dental group with three teams outside the clinic: front office, billing and scheduling. The operations director noticed that the front office used AI freely, billing used it in private, and scheduling avoided it.

Illustrative example.

  • Front office.The team’s question was what is allowed. Several people had pasted patient details into an assistant to draft letters. The page now says AI drafts general replies and letters from a template, and patient details stay in the practice system.
  • Billing.The team’s question was whether its way was right. Each coordinator drafted claim follow-ups differently and redid them by hand. The billing lead now owns one shared workflow for follow-ups, and the senior coordinator checks each one before it goes out.
  • Scheduling.The team’s question was what AI meant for its roles. The page now says the workflow drafts the week’s reminder and rebooking messages, and each scheduler decides which patients to call in person and keeps the call list.

The director read each page to its team. Billing asked to add a line on denied claims. Scheduling asked who to tell about a better reminder they had found. Both went on the page.

What is different in a company of under 200 people.

In a company of under 200 people, the leader answers the three questions in person. A large company sends a memo and a training course. You can read each team its page and hear its questions the same week.

  • Unease spreads faster.Small teams talk. One person’s worry becomes the team’s within days, and one clear answer settles it as quickly.
  • One role is often one person.When the plan describes a role, the person reads it about themselves. Write it with them.
  • The lead does the work too.A team lead who also runs billing has the same questions as the team. Answer the lead’s first.

What do we tell employees about AI? gives the announcement that goes with the pages.

Write the three answers for each team, and the unease has somewhere to go. FullyPowered Planner holds the plan, team by team and role by role, and your colleagues add their own teams. How organizations plan their use of AI covers the rest of the method.