adopt-team-involvement-plan git:20261026.a2f0899A

Immutable. This exact content never changes and is served at /api/v1/blob/175be10023cde8da.

---
name: adopt-team-involvement-plan
description: Builds a team involvement plan for an AI rollout with what is decided, open and unknown, the level of influence promised for each open choice, three concrete questions, and a ledger of what was heard, changed and declined. Use for "run adopt-team-involvement-plan", "team involvement plan", "involve people in the AI rollout", "consult the team about AI", "what can people still change", "how do we get real input on the AI plan", "people feel the decision was already made", part of the Claude Guide for AI Adoption Pack by 90n.
---
# Team Involvement Plan
## When To Use
The rollout is "consulted" after the decision and people can tell. You are planning how the people affected by an AI change get a say across the programme, and you need to be honest about how much say that is. The question this answers: what can people still influence, how do they do it, and how will they hear what happened to what they said?

## When Not To Use
Not for a single session where a leader invites objections before one named decision: use Team Challenge Invitation. Not for following up concerns already raised one by one: that is the Concern Response Log. If nothing at all is open, skip consultation and send a plain announcement.

## Inputs
- The change and its status (what is already decided, by whom, what is still open) and the groups affected, including people who receive the output of the work.
- The decision owner and dates, any formal consultation route that exists, and access needs (shifts, remote staff, languages).
If you have none of this, I start from one sentence on the change and who decides, and mark the output as a first draft.

## Approach
The plan uses the spectrum of public participation, designed by the International Association for Public Participation and described in a public guide from the US Environmental Protection Agency: five levels of influence, each carrying a promise to the people asked. The judgement is choosing the level you can keep, choice by choice. The failure it prevents is the workshop with sticky notes held a week after the licences were signed: people give their time, nothing can move, and the next request for input gets silence.

## Workflow
1. Ask at most three questions: what is already decided and by whom, which choices are still open and until when, and who is affected beyond the obvious team.
2. Build the decided, open and unknown table first. Be strict: a choice the sponsor will not reverse is decided, whatever the slide says. If nothing is open, call the activity implementation feedback, not consultation.
3. For each open choice, pick one level and write its promise in full: inform (we keep you informed), consult (we consider your input and report back how it influenced the decision), involve (you have access to the process and get direct feedback on how your input shaped the outcome), collaborate (your input is built in as far as possible), hand over the decision (we implement what you decide). Pick the lowest level you are certain to keep, never the one that sounds best.
4. List who is affected: people doing the work, people receiving its output, people carrying extra work during the change. Check the less visible: other shifts, contractors, remote staff. Ask whose experience is missing.
5. Draft three concrete questions about a real piece of work: will this work on [task], what would you lose, what would you do instead. Questions about how people feel about AI in general produce nothing a decision can use.
6. Set the channels: in paid time, in accessible formats, with a private route through an appropriate person. Do not promise anonymity a small group cannot keep. Set the date input closes, before the decision date, and name who reads the input. Then set up the ledger: what was heard, what changed, what was declined and why, returned by a named person on a set date. Where a formal consultation duty may apply, flag it: check with a qualified adviser.

## Output Format
```markdown
# Team Involvement Plan
**Change:** [what] · **Decision owner:** [name] · **Input closes:** [date] · **Decision:** [date] · Draft status: [first draft / agreed on date]
## Decided, open, unknown, and the level of influence
| Topic | Status | Who decided or will decide, by when | Level (open choices only) | The promise we make, and who keeps it |
|---|---|---|---|---|
| [choice] | [decided / open / unknown] | [name, date] | [inform / consult / involve / collaborate / hand over the decision] | [promise in one sentence, name] |

## Who is affected
| Group (role, never names) | How the change touches their work | Less visible? |
|---|---|---|
| [group] | [doing the work / receiving its output / carrying extra work] | [other shift, contractor, remote] |

## The three questions
1. Will this work on [real task]? 2. What would you lose? 3. What would you do instead?
## Channels and access
| Channel | When, in paid time | Format and access | Private route |
|---|---|---|---|
| [session, written, one to one] | [date and slot] | [format] | [appropriate person, and what cannot be kept anonymous] |

## Ledger (returned by [name] on [date])
| What was heard | Changed | Declined, and why | Still open |
|---|---|---|---|
| [input, with no name attached] | [what changed] | [reason] | [next date] |

## Decision
[Decision owner] confirms what is open and the promise for each choice by [date], before any invitation goes out, and returns the ledger to everyone asked by [date].
```

## Done When
- Every open choice has one level and a promise somebody named has agreed to keep.
- Input closes before the decision date, with time left to act on it.
- The affected list covers less visible groups, and the ledger has an owner, a return date and a column for what was declined.

## Quality Bar
- Nothing already decided is presented as open, and the plan says so in plain words.
- Questions are about a real piece of work, never about attitude to AI.
- Attendance is not consent, silence is not agreement, and nobody is scored, listed or noted for taking part, staying away or disagreeing; the ledger records input, never who gave it.
- A person runs every conversation, in paid time; Claude drafts the plan and the questions.
- Only promise the influence people really have; a named person returns what was heard, changed and declined.

## Next
Run adopt-capacity-check (Adoption Capacity Check) to see whether people have the time the plan assumes.

## 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).