You're offline, showing the last version we saved.
Free template

30-day AI training plan tracker

The 30-day plan works when someone can see who is where. This workbook holds the plan from the playbook, a row for every person with a status each week, the firm tasks and where their prompts are saved, and the numbers the week-four review needs.

Updated · Excel workbook (.xlsx) · Free with a firm profile

Free with a firm profile

Get the editable 30-day AI training tracker

The Excel workbook (.xlsx) is free with a firm profile. Enter your email and we will send a code to unlock it. No password, and one profile unlocks every template.

Have a password? Sign in

By continuing you agree to our terms and privacy policy.

Or build it with AI

Open Claude or ChatGPT with the prompt filled in. It builds the workbook for your firm straight away, then asks a few questions to tailor it. No client data needed.

Open in ClaudeOpen in ChatGPT

Works without client data. Don't paste client names, tax file numbers or ledgers into an AI tool your firm hasn't approved for them. AI data safety playbook.

How to use it

How to use this template

  1. Name an owner for each stage

    On Plan, put one name and a due date against each stage. The owner answers for AI in the firm; champions run the desk time. The plan itself is in the 30-day AI training plan.

  2. List the first cohort

    On Team progress, one row per person, with your one or two champions marked. Update the weekly status as each stage finishes.

  3. Record the firm tasks

    On Firm tasks, the three tasks from week two and the role tasks from week three, with who owns each prompt and where it is saved.

  4. Run the week-four review from the numbers

    Fill in seats issued and active from your tool’s admin view, and the near misses reported. Then record the decisions the review ends with.

The template

The template in full

30-day AI training plan tracker

Replace everything in [square brackets].

What's in the workbook

Four sheets after the read me: the plan, the team’s progress, the firm tasks, and the week-four review.

SheetWhat it does
Read meWhat the workbook is for, how to use it, and where it came from.
PlanThe 30-day plan from the playbook, one row per stage, with an owner, a due date and a status for each.
Team progressOne row per person: their status each week, whether they have confirmed the policy, how many of the firm tasks they ran last week, and their % complete.
Firm tasksThe three firm tasks from week two and the role tasks from week three: who owns each prompt, where it is saved and whether it has been tested on live work.
Week 4 reviewThe numbers for the week-four review, and the decisions it should end with.

Shaded cells are yours to fill in. Everything else calculates from them. Nothing is filled in for you: no benchmark, rate, salary or fee is typed in as a default, so every result is built from your own numbers.

Sheet 1: Plan

The 30-day plan from the playbook, one row per stage, with an owner, a due date and a status for each.

DaysFocusSessionHomeworkDone when
Before day 1Policy, seats, championsOwner and champions, 60 minChampions try the tool on internal workPolicy signed, seats live
Days 1 to 7Safe use and the policyWhole team, 45 minOne safe internal task in the firm accountEveryone confirmed the policy in writing
Days 8 to 14Three firm tasksWhole team, 30 minRun all three tasks on live jobsThree tested prompts saved and shared
Days 15 to 21Role tracks and reviewPer role, 30 min eachOne role task; reviewers check AI work against sourceEach role has an added task
Days 22 to 30Measure, fix, decideWhole team, 30 minTell the champion what you tried that failedUsage numbers, near misses, policy revised
Every weekChampion desk timeChampion, 60 min openBring one stuck taskQuestions logged and added to the FAQ
ColumnTypeWhat it is
DaysSet for youWhen it happens.
FocusSet for youWhat the stage is about.
SessionSet for youWho meets, and for how long.
HomeworkSet for youWhat everyone does before the next stage.
Done whenSet for youHow you know the stage is finished.
OwnerYou fill inOne named person per stage.
Due dateYou fill inWhen the stage should be done.
StatusPick from a list: Not started, In progress, DoneWhere the stage is up to.
  • Bottom rows (Stages done): totals and averages across the rows you have filled in.

Sheet 2: Team progress

One row per person: their status each week, whether they have confirmed the policy, how many of the firm tasks they ran last week, and their % complete.

ColumnTypeWhat it is
PersonYou fill inEveryone in the first cohort.
RoleYou fill inBookkeeper, accountant, admin, partner.
Champion?Pick from a list: Yes, NoMark your one or two champions.
Week 1: safe use and policyPick from a list: Not started, In progress, DoneThe 45-minute session and one safe task.
Week 2: three firm tasksPick from a list: Not started, In progress, DoneAll three tested prompts run on live jobs.
Week 3: role trackPick from a list: Not started, In progress, DoneThe role session and one role task.
Week 4: measure and reviewPick from a list: Not started, In progress, DoneThe review session, and what failed reported.
Policy confirmed in writing?Pick from a list: Yes, NoSigned or acknowledged the AI policy.
Firm tasks run last week (0 to 3)You fill inHow many of the three firm tasks they ran on live work.
% completeCalculatedWeeks marked Done / 4.
NotesYou fill inAnything stuck.
  • % complete: Weeks marked Done / 4.
  • Bottom rows (Team): totals and averages across the rows you have filled in.

Sheet 3: Firm tasks

The three firm tasks from week two and the role tasks from week three: who owns each prompt, where it is saved and whether it has been tested on live work.

ColumnTypeWhat it is
TaskYou fill inA job the firm does every week.
KindPick from a list: Firm task, Role taskFirm tasks are for everyone; role tasks for one role.
RoleYou fill inFor role tasks: which role.
Prompt ownerYou fill inThe champion who wrote and tested it.
Where the prompt is savedYou fill inThe shared project, saved instruction or skill.
Tested on live work?Pick from a list: Yes, NoRun on a real job, not a made-up one.
StatusPick from a list: Not started, In progress, DoneWhere it is up to.

Sheet 4: Week 4 review

The numbers for the week-four review, and the decisions it should end with.

LineTypeHow it works
Seats issuedYou fill inBusiness-plan seats the firm set up.
Seats active in the past weekYou fill inFrom your tool's admin view. Check what yours shows rather than assuming.
Share of seats in useCalculatedActive seats / seats issued.
Firm tasks run last week, whole teamCalculatedFrom Team progress.
Team % completeCalculatedAverage of everyone’s % complete.
Near misses reported this monthYou fill inA near miss reported is a good sign: people trust the rule.
Task that saved the most timeYou fill inKeep it.
Task to retireYou fill inReplace it with one new one.
Policy gap foundYou fill inUpdate the policy before the next cohort.
Next cohortYou fill inThe decision the month ends with.
  • Share of seats in use: Active seats / seats issued.
  • Firm tasks run last week, whole team: From Team progress.
  • Team % complete: Average of everyone’s % complete.

Before you rely on it

If the month lands in a busy season, stretch the plan to six weeks rather than skip the safe-use week. The tracker works the same either way: the statuses matter, not the dates.

This workbook is a starting point, not financial, tax, employment or legal advice. It only knows the numbers you give it, so check them against your own records and take advice on anything you are unsure of.

Common questions

Frequently asked questions

How do you track AI training across an accounting team?
One row per person with a status for each week of the plan, whether they have confirmed the AI policy in writing, and how many of the firm’s shared AI tasks they ran on live work last week. The tracker works out each person’s % complete and the team average.
What should the week-four review measure?
The share of seats in use (from your tool’s admin view), how many firm tasks the team ran in the past week, and how many near misses were reported. Then decide which task to keep, which to retire, what the policy missed and who gets seats next.
Why count near misses?
A near miss reported means people know the rule and trust that reporting is not punished. A month with none usually means nobody is reporting, not that nothing happened. The AI data safety playbook defines what counts as a near miss.
Does the tracker work with any AI tool?
Yes. It tracks people, tasks and statuses, not a particular product. Record where each prompt is saved in whatever form your tool offers: a shared project, a saved instruction or a skill. Claude, ChatGPT and Copilot compared helps you choose which to train on.
More templates

More templates

The thinking behind these: AI for accountants and bookkeepers: what actually works.

Read the notes →

The AI in Practice Summit notes are up: a quick read, the full notes and a transcript for every session, plus the replays and Ask the Summit. Free.