Marcus Andrey Study
Ler em português Library

Framework · AI & design · 3 min read

AI champions network

A small, rotating group of designers (one per 3–5 colleagues) who try tools, write playbook entries and run office hours, so the whole team learns from a few people's experiments.

1Try tools

Test approved tools on real, low-risk work and note what helps.

Typically: one new use case a month
2Write playbook

Turn each experiment into a playbook entry others can repeat.

Typically: task, tool, prompt, pitfalls
3Office hours

Hold a regular open slot where anyone brings a task or question.

Typically: 30 minutes, every week
4Rotate

Hand the role to new people so knowledge spreads across the team.

Typically: every quarter

One champion per 3–5 designers · enthusiasts and at least one sceptic

An AI champion has four jobs, and writing things down is the one that makes the others count.

Use it when

  • A few designers are already experimenting alone and nobody else benefits from what they learn.
  • You are starting the pilot phase of a 90-day AI plan and need people to run pilots and share results.
  • The team's prompts, specs and tips are scattered across chats and personal notes.

How to use it

  1. Size the network. Plan one champion per 3–5 designers. A team of 8 needs two; a team of 20 needs four to six. Output: the number of champions and the groups they support.
  2. Pick a mix, not just fans. Choose enthusiasts who are already experimenting and at least one thoughtful sceptic who cares about quality, ethics or data. Ask; don't assign it as a reward. Output: named champions who said yes.
  3. Give them time and a clear remit. Agree protected time, for example 10% in the first quarter, and write down the three jobs: try tools, write playbook entries, run office hours. Make clear they work inside the AI usage guidelines, not around them. Output: a one-paragraph remit agreed with each champion and their manager.
  4. Set up the shared places. Create the AI playbook (one wiki page per use case: task, tool, prompt or spec template, review notes, time saved, pitfalls) and the prompt library the champions curate. Book a recurring office hour or show and tell. Output: a playbook page, a prompt library and a calendar slot.
  5. Review and rotate. Each quarter, look at what was shared and what people used. Then rotate at least half of the champions so new people learn and the first group does not burn out. Output: a short quarterly note and the next set of champions.

Worked example

An internal operations platform with 6 product teams and 8 designers. The UX lead picks two champions: an enthusiast who already builds code prototypes, and a senior designer who has raised concerns about hallucinated research quotes.

Champion A (enthusiast) Champion B (sceptic)
Supports 4 designers in the order and billing teams 4 designers in the warehouse and reporting teams
Time 10% of their week this quarter 10% of their week this quarter
First playbook entries "Spec to code prototype with design-system components" "First-pass research clustering, and how to check every quote"
Office hours Tuesday, 30 minutes, open to all Shares the same slot every other week
Result after a quarter 6 playbook entries, 12 prompts in the library, two designers now prototype in code The research entry became part of the team's review checklist

Next quarter, Champion B stays for continuity and a pragmatist from the reporting team replaces Champion A.

Watch out for

  • Only picking enthusiasts. A network of fans produces hype and misses risks. The sceptic is the one who makes the playbook trustworthy.
  • No real time. "Champion on top of a full sprint" means nothing gets written down. Protect the time with their manager.
  • Champions as gatekeepers. If people need a champion's permission to try something, the network slows learning. Champions share and guide; the guidelines set the rules.
  • Never rotating. The same two people become the "AI people", everyone else waits for them, and they burn out.

Template

Copy this into your team wiki and fill one block per champion.

AI champion:                      Term: Q_ (rotate after)
Supports (3–5 designers):
Protected time:                   Agreed with:

Remit
- Try tools (approved list only):
- Playbook entries this term (target):
- Office hours (day, time, length):
- Prompt library: curate and clean up

Playbook entry format
Task | Tool | Prompt or spec template | Review notes | Time saved | Pitfalls

End-of-term review
Shared:            Used by others:            Hand over to:

"See on Amazon" links are affiliate links. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.

Saved only on this device.

↑ ↓ to move · Enter to open · Esc to close