The short answer.
New hires learn the AI workflows their role runs by running them, with a person checking. On day one a new person gets the team’s best method for each piece of work written down, a first draft the workflow has prepared, and one named colleague who reviews how they checked it. They learn the job by reviewing prepared work against the team’s standard, so they start strong and become productive sooner.
The company’s plan tells them which workflows come with the role. The plan names the work, the workflow carries the method, and a person teaches the judgment.
The workflow is the training.
An AI workflow is your team’s best way of doing one piece of work, written down precisely enough that an AI assistant prepares it the same way every time. That same written method is the clearest training a new hire will ever get. It says what the work is for, which facts it needs, what good output looks like and where a person decides.
A new hire trained on a workflow learns in three layers.
- The method.They read how the team does the work: the steps, the standard and the exceptions the team has already met.
- The output.They see what a good draft looks like, because the workflow produces one in front of them on real work.
- The judgment.They review the draft, decide what to change and send it on. A named colleague reviews that review until the new person’s checking matches the team’s standard.
The workflow prepares; the new person decides. Deciding takes a person to teach, and the workflow frees the experienced colleague to teach it.
The plan says which workflows come with the role.
A company’s AI plan lists its teams, the roles in each team and the workflows each role runs, so the plan doubles as the new hire’s list of what to learn. A new account coordinator opens the plan, finds their team and role, and sees the four or five workflows that come with the job, each with its owner.
Without a plan, a manager reconstructs that list from memory in the first week, and the new person learns whatever the nearest colleague happens to use. How to manage AI across teams without everyone doing their own thing covers the same problem at company scale. In FullyPowered Planner, the role already carries its workflows, so the list is waiting before the new person arrives.
How AI changes onboarding.
AI changes onboarding by giving a new hire real, prepared work to review from the first day, instead of weeks of shadowing before they touch anything. The experienced colleague stops being the only source of the method and becomes the reviewer of the new person’s judgment.
- Day one has real work in it.The workflow prepares a draft on a live case. The new person reviews it with the method beside them.
- The standard is visible.The new person compares each draft with the team’s written standard, so “good” is on the page, not in someone’s head.
- Questions get sharper.A new hire who has reviewed ten drafts asks about the exceptions, which are the questions worth an experienced colleague’s time.
- Review steps down on purpose.The reviewer checks every piece at first, then a sample, then the exceptions, as the new person’s checking matches the standard.
How you stop a new person learning five ways to do one job.
You stop a new person learning five ways to do one job by giving every role one workflow per piece of work, owned by one person who keeps it current. When five colleagues each prompt an assistant their own way, a new hire inherits all five habits and none of the reasons. When the team runs one workflow, the new hire learns the method the team agreed.
Three things hold the single method in place.
- One owner per workflow, by name.The owner decides what changes and tells the team when it does.
- One current version.Everyone in the role runs the same version, in the company’s own ChatGPT, Claude, Copilot or Gemini account.
- A place for disagreements.When an experienced person does it differently and better, that becomes a change to the workflow, and every new hire after gets it.
Turning one experienced person’s way of working into the team’s method is its own piece of work. From one person’s know-how to a method the whole team runs shows how it is done.
A first-week sheet to copy.
Fill in one first-week sheet for each new person in a role before they start. Copy the headings below.
- Role and team.The role’s name, the team it sits in and the person the new hire reports to.
- The workflows that come with the role.Each workflow by name, as it appears in the plan.
- Who owns each workflow.One person per workflow, by name.
- What to review first.The one workflow to start on, and the first three real cases to review with it.
- Who reviews the new person’s work.One colleague, by name, and how much they check: everything, a sample or exceptions only.
- When review steps down.What the new person’s checking has to show before the reviewer moves from everything to a sample.
- What to ask in week one.Which cases does this workflow get wrong? Where do I decide, and where do I pass it on? Who do I tell when something in the method looks out of date?
One company’s first week for a new coordinator.
Illustrative example.
Take a 120-person property management company hiring a second maintenance coordinator. The company’s plan lists the maintenance team’s coordinator role with four workflows: triaging tenant repair requests, preparing work orders for contractors, drafting tenant updates, and preparing the weekly open-jobs summary for the operations lead.
Before the start date, the operations lead fills in the first-week sheet. Triage is the workflow to start on, because every other workflow begins with it. The experienced coordinator owns triage and work orders and is named as the new person’s reviewer.
On day one the new coordinator reads the triage method, then reviews the drafts the workflow prepares on that morning’s real requests. The experienced coordinator checks each of the new person’s decisions and explains the ones they would have made differently. Two of those explanations turn out to be rules nobody had written down, and the owner adds them to the workflow, so the next hire gets them too. By the end of the week the new coordinator is reviewing work orders as well, and the reviewer has moved to checking a sample.
What is different in a company of under 200 people.
In a company of under 200 people, the workflow often replaces a training program that never existed. Most published advice on onboarding assumes a learning team, a course library and a cohort of hires. A smaller company hires one person at a time into a role, and the training has always been the colleague at the next desk.
- The experienced person is one person.When they are the only source of the method, every new hire takes their time. The written workflow gives that time back for the judgment that needs them.
- Every hire matters.One new person in a team of six is a large share of the team’s work. Starting them on real, prepared work matters more here than anywhere.
- The plan is small enough to keep current.You can name every role and its workflows on one page, so the new hire’s list is always right.
The workflow map for the people function shows the recruiting, training and development work a small people team can hand to workflows. Who should own AI in a company your size? covers who keeps the plan and its roles current.
Where to start.
Start with the plan, because the plan is what tells a new person which workflows come with their role. How organizations plan their use of AI walks through it in three steps: plan it, develop the workflows, deploy them. How to deploy AI workflows to a whole team, and keep them working covers the rollout every new hire later joins. FullyPowered Planner is where the plan lives: you set out your teams and their roles, and each role carries the workflows a new person will learn.