---
name: adopt-manager-ai-kit
description: Builds a one-page kit for each manager with what to say in the first team conversation, one piece of their own use to show, three questions for meetings that already exist, what to remove or protect, and what a manager never does. Use for "run adopt-manager-ai-kit", "manager AI kit", "what should managers say about AI", "managers never mention AI", "brief managers on the AI rollout", "manager talking points for AI", "how managers support AI use", "equip managers for AI adoption", part of the Claude Guide for AI Adoption Pack by 90n.
---

# Manager AI Kit
## When To Use
The tool is live, the training ran, and the manager never mentions AI, or mentions it only as a target. Use this when you equip many managers and need to hand each one a page they can act on this week. It answers one question: what does a manager say, show, ask and remove, as a standing part of the job, so the team feels free to try?

## When Not To Use
If the need is the before and after around one named session or course, run Manager Follow-Up Guide: this kit is tied to no date and carries no conversation with one person. If the team has to agree what is fine, run Team AI Agreement; if the worry is where saved time goes, run Freed Time Agreement.

## Inputs
- The work change expected of the teams, per role, and where the organisation's AI rules live.
- What managers have been told so far (any target word for word), the meetings they already run with their teams, and the obstacles they already report.
If you have none of this, I start from one manager role and one team task and mark the output as a first draft.

## Approach
In Gallup's 2026 workplace surveys, frequent use goes with a manager who actively supports AI use, fit with existing work, room to experiment and clear policies. Transfer of training research (Blume, Ford, Baldwin and Huang, 2010) points the same way: use back at work relates to a supportive work environment. Both are correlational, and the kit says so and promises nothing. The judgment is that a manager's part is small and visible: a few sentences, one honest example, three questions, one obstacle gone. The failure it prevents: the manager forwards the launch email, says nothing for three months, then asks someone "are you using it yet?", and the team learns that AI is something you get checked on.

## Workflow
1. Ask three questions: what should people in these teams do differently in their work, what have managers been told (any target word for word), and which meetings do they already hold?
2. Say. Draft a short script for the first team conversation in words a manager would use: why this for our work, what is and is not expected, that trying it and not using it are both fine to talk about, where the rules are, and that nobody's use is counted. If a target exists, the script names it as a team topic and never as a personal one.
3. Show. Leave a blank for one piece of the manager's own use on real work, with what did not work and what they checked. A manager who does not use it yet says so and learns alongside the team; a staged success teaches people to hide their failed tries.
4. Ask. Write three questions for meetings that already exist, put to the team about the work: what did we try, what did we stop doing, which rule is unclear. None is addressed to one person, and none asks who.
5. Remove or protect. One obstacle from each of time (protected practice), access (a licence, the tool) and rules (one unclear rule to get answered), each with what the manager will do and by when. If the manager cannot remove it, the kit says who can.
6. Never. State the four lines in full: count, compare or report who uses it; set a personal usage target; raise AI use in a performance review; ask to see anyone's prompts. Then fit the kit on one page per manager, with blanks they fill in their own words, and check that no field asks the manager to record anything about a named person.

## Output Format
```markdown
# Manager AI Kit
For: [manager role] | Team: [team] | Prepared by: [enablement lead] | Version: [date]
## Say: the first team conversation
| Point | In your own words |
|---|---|
| Why this, for our work | [script line] |
| What is expected, and what is not | [script line] |
| Trying it and not using it are both fine to talk about | [script line] |
| Where the rules are | [location, and who to ask] |
| Nobody's use is counted | [script line] |

## Show: one piece of my own use
Task: [real task] | What I gave it: [input] | What did not work: [what happened] | What I checked: [check] | If I do not use it yet: [how I say so, and what I try alongside the team]
## Ask: three questions for meetings that already exist
In [meeting], put to the team, about the work: (1) What did we try? (2) What did we stop doing? (3) Which rule is unclear?
## Remove or protect
| Obstacle | What I will do | Who else is needed | By when |
|---|---|---|---|
| Time | [practice time protected] | [role] | [date] |
| Access | [licence or tool fix] | [role] | [date] |
| Rules | [one unclear rule to get answered] | [policy owner] | [date] |

## A manager never
Counts, compares or reports who uses AI. Sets a personal usage target. Raises AI use in a performance review. Asks to see anyone's prompts.
## Decision
[manager] holds the first team conversation by [date]; [enablement lead] and [sponsor] decide how the obstacles managers raise get fixed, and by whom, by [date].
```

## Done When
- The kit fits on one page per manager and every script line is something a manager would say out loud.
- The show section has room for what did not work, and the three questions are put to the team about the work, in meetings that already exist.
- Time, access and rules each have one obstacle, an action and a date.

## Quality Bar
- The survey and transfer evidence is described as correlational, with no figure and no promised result.
- The kit is the standing role: no session date, no conversation with one person, no first task, no log.
- No field records anything about a named person; Claude drafts the script, the manager rewrites it and says it.
- A manager supports and removes obstacles; a manager never counts, compares or reports who uses AI.

## Next
Run adopt-team-ai-agreement (Team AI Agreement) to have the team agree what is fine instead of guessing.

## About the makers
This pack is made by 90n ([90n.ai](https://90n.ai)): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).