---
name: adopt-capacity-check
description: Prepares an adoption capacity check with the learning and practice time the change needs per role beside current delivery load, the gap, what can pause, shrink or move, and options with trade-offs and prerequisites for starting. Use for "run adopt-capacity-check", "adoption capacity check", "we have no time to learn AI", "find time for AI training", "do people have capacity for the AI rollout", "protected practice time", "learning AI on top of a full week", part of the Claude Guide for AI Adoption Pack by 90n.
---
# Adoption Capacity Check
## When To Use
People are told to learn AI on top of a full week, and the programme assumes goodwill makes hours. You have a plan with sessions and practice in it and no line saying where the time comes from. The question this answers: how much time does this change really need per role, how much exists, and what gives?

## When Not To Use
Not when you do not yet know that time is the obstacle: run AI Adoption Barrier Diagnosis first, since the block may be access or an unclear rule. Not for deciding what happens to time once tasks get lighter: that is the Freed Time Agreement.

## Inputs
- The change and the roles it touches, with what each role has to learn and practise.
- The planned sessions, practice and support, and any period of running old and new ways side by side.
- Current delivery load per role (commitments, deadlines, peaks, cover), who can authorise moving work, and the sponsor.
If you have none of this, I start from the roles and your own rough estimate of the time needed, and mark the output as a first draft.

## Approach
The check is organised by demands, control and support, three of the six areas in the Management Standards of the UK Health and Safety Executive. They were written for workload and stress at work and are borrowed here for one narrow question. An official report on a public sector trial of an AI assistant noted that some staff lacked the capacity to take part in training because of their workload. The failure this prevents: the session lands in a deadline week, practice is left to evenings, and the programme then reads the empty room as a lack of interest.

## Workflow
1. Ask at most three questions: what does each role have to learn and practise and over what period, what does the current load look like in those weeks, and who can agree to pause or move work.
2. Demands. For each role, set the time the change needs beside current delivery load, split into sessions, practice on real work, running old and new ways side by side, and support. Use ranges you give. Every number is your assumption and is labelled as one; I invent none. Then show the gap per role, even when it is uncomfortable: goodwill does not create hours. If the estimate is too rough to trust, say what would firm it up, such as one team noting a real week.
3. Control. List what people can decide about when and how they practise, then what can pause, shrink or move, with who can authorise each item. "Deprioritise" with no name beside it is not a decision.
4. Support. Name who helps (a trainer, a peer, the manager), when, and whether that person has the time. Check the managers' own load: they are asked to support the change while learning it themselves.
5. Build the options with their trade-offs: smaller scope, later start, protected practice time, extra cover, or stop. For each, what it costs, what it protects and who can authorise it.
6. Draft a conditional start: named prerequisites, an owner for each, a review date. Objections that are not resolved stay on the record. The sponsor chooses.

## Output Format
```markdown
# Adoption Capacity Check
**Change:** [what] · **Period:** [start] to [end] · **Sponsor:** [name] · All figures are assumptions supplied by [name] on [date], shown as ranges.
## Demands: time needed and current load, per role
| Role | Sessions | Practice on real work | Old and new side by side | Support | Current delivery load | Gap |
|---|---|---|---|---|---|---|
| [role] | [range] | [range] | [range] | [range] | [commitments and peaks] | [range, or "none found"] |

## Control: what can pause, shrink or move
| Work item | Pause, shrink or move | Who can authorise | What people decide themselves (when, how) |
|---|---|---|---|
| [item] | [choice and until when] | [name] | [choice left to the person or team] |

## Support
[Trainer, peer or manager] helps with [what] in [slot]; has the time: [yes / no / not known, with the reason]. Managers' own capacity: [stated].
## Options
| Option | What it costs | What it protects | Who can authorise |
|---|---|---|---|
| [smaller scope / later start / protected practice time / extra cover / stop] | [trade-off] | [what it makes possible] | [name] |

## Prerequisites for starting
1. [Condition] · owner [name] · due [date] · [open / met]

Objections and open questions on record: [list, with no names attached].
## Decision
[Sponsor name] chooses one option and confirms the prerequisites by [date]; the check is reviewed on [date] against what the weeks really looked like.
```

## Done When
- Every role has a time needed, a current load and a gap, each marked as an assumption with its source.
- Each thing that pauses, shrinks or moves has a name beside it who can authorise it.
- Managers' own capacity is stated, and the options include stopping or starting later, not only ways to squeeze the time in.

## Quality Bar
- Roles, never names. No table lists a person's hours, pace or availability.
- No readiness score and no pass mark, for a person or for a team.
- An unresolved gap is shown as a gap, never closed with "people will find the time"; doubts about the change itself stay on the record, apart from the time question.
- Working time and health at work questions: check with a qualified adviser.
- Capacity is a fact about the work, not about willingness; nobody is asked to learn in unpaid time, and the sponsor chooses the option.

## Next
Run adopt-ai-basics-curriculum (AI Basics Curriculum) to design learning that fits the time that really exists.

## 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).