AI / Me / We
AI-HUMAN ROLESENABLE
Lauren Kelly
Purpose Draw one clear map that defines role boundaries: AI, Human, Shared.
Intervention type Role-clarity workshop
Audience Leaders and managers who set AI goals, run roll-outs, and own adoption KPIs
Time 75 min live session · 2 min “drift” check every sprint retro
Expected outcomes
Users: Is clear on “AI does ___, I do ___”.
Teams: A3 (or Miro) Role Canvas posted; one shared KPI pinned on top; drift ritual on calendar.
Business: Fewer hand-off errors, faster cycle times on mapped tasks (target: ≥10 % within 4 weeks).
Organisation: Anxiety about “AI stealing jobs” and loss of instigation knowledge falls.
What to bring to the session
AI / Me / We Canvas. (Build below)
Task Deck. 30 common workflow cards (below)
Scenario Cards. (below)
Steps
1 | Setup
Step 1: Prep your board
Use a whiteboard, wall, or table. Lay out your AI / Me / We Canvas with three columns:
AI = work the tool can do independently
Me = tasks that need human judgment or oversight
We = shared ownership (e.g., AI draft, human edit)
Draw or print the canvas with the KPI bar at the top. Make sure you have enough space for 10–15 task stickies.
Step 2: Get your cards ready
Shuffle the Task Deck and lay the cards in the centre.
Choose 1–2 Scenario Cards in advance — pick ones that match current stressors (e.g. “volume spike”, “policy change”, “tech outage”).
Have the KPI Cheat-Sheet printed or open on screen for quick reference.
2 | Draft the Tasks
Step 3: Build your list
Each person picks up to 12 tasks using the Task Deck or their real workflow.
Stick these beside the canvas.
3 | Sort the Work
Step 4: Drop each task into AI / Me / We
AI → “Can the tool run this unsupervised?”
Me → “If AI fails here, who owns the risk?”
We → “Do we need both for speed, quality, or judgment?”
Discuss. Debate. Decide.
Tip: Add notes as to how and why for later.
4 | Stress-Test the Boundary
Step 5: Add pressure, see what moves
Pick one Scenario Card and ask aloud:
What breaks first?
What shifts to humans under pressure?
What needs stronger guard-rails?
Tip: Document your conversations for the next task.
Step 6: Add mitigation ideas to critical tasks
Look at tasks that raised concerns or shifted under pressure.
Ask:
What could go wrong here, and why?
What would reduce that risk? (Process, prompt, alert, policy…)
Who should own that mitigation? AI, human, or both?
Write the mitigation beside the sticky using a small post-it or note box.
Label the owner (AI, Me, We).
Example:
Task = “Initial draft of policy update”
Concern = “Regulatory risk if misunderstood”
Mitigation = “Human legal check before send → Me”
5 | Lock the Shared KPI
Step 7: Choose one metric that proves this setup works
Use the KPI Cheat-Sheet to find a task-relevant metric (e.g., “Turnaround time < 24h while maintaining Quality ≥ 8/10”).
Resources
Task Deck
40 common workflow uses. Use them as a starting point to your discussions.
Scenario Cards
Use these to pressure-test your role boundaries.
Example KPIs
Other methods within the flow block
Human-AI Performance
By Lauren Kelly
Contact: lauren@alterkind.com
© 2025 Alterkind Ltd. All rights reserved.
Human-AI Performance™ is a proprietary methodology developed by Alterkind Ltd using our Behaviour Thinking® framework. All content, tools, systems, and resources presented on this site are the exclusive intellectual property of Alterkind Ltd.
You’re welcome to use, share, and adapt these materials for personal learning and non-commercial team use.
For any commercial use, redistribution, or integration into client work, services, or paid products, please contact lauren@alterkind.com to discuss licensing term
Icons by Creative Mahira, The Noun Project.
Thanks to Nicholas Edell, Valentina Tan and multiple VPs implementing AI for your feedback during development.
LICENSE
Human AI Performance by Alterkind is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Based on work at alterkind.com
For commercial licensing contact: lauren@alterkind.com