Drift Signals Scan
AI RESILIENCEADAPTBEHAVIOUR KIT
Lauren Kelly
Purpose Detect early signs of behavioural slippage
Intervention type Behavioural stewardship & embedded accountability
Audience Team leads, product managers, delivery owners, or ops leads responsible for day-to-day AI performance
Time 1 hour setup, 1 hour fortnightly check-in, optional 60-min monthly review ≈ 4h/month
Expected outcomes
Users: Know who to go to when the AI behaviour breaks, drifts, or needs adjusting
Teams: Build a clear review rhythm and behavioural playbook, led by a trusted peer
Business: Avoid costly drop-offs post-rollout with visible accountability and reduced friction
Organisation: Behavioural adoption becomes owned and managed at team level.
What to bring to the session
HAP Behaviour Kit. (The Pattern cards)
Task Target Sheet. (Below)
Drift Types. (Below)
Drift Signals Grid. (Below)
Steps
1 | Choose a Behaviour to Scan
Step 1: Pin down the task and behaviour you’re scanning for drift
Use the Task Target Sheet to guide this.
Ask:
What’s the task or moment where AI is meant to be used?
What’s the intended behaviour? (Be specific)
Use this sentence starter:
“We expect team members to [do what?] using AI during [which task or moment?].”
For example: “We expect everyone to use AI to generate a first-draft report before editing manually.”
Behaviour checklist:
Is it observable?
Is it repeatable?
Is there a known trigger or moment?
If not, pause and clarify.
2 | Spot Behavioural Drift
Step 2: Look for early warning signs of slippage
Use the Drift Signals to structure your conversations and exploration.
Ask:
“Where is this behaviour starting to shift, fade, or get skipped?”
Tip: Use the Drift Types (below) to guide your observations.
3 | Use HAP Patterns as Drift Mirrors
Step 3: Scan for pre-identified behavioural friction patterns
Bring 3–5 relevant patterns from the HAP Behaviour Kit (e.g. “Thinks AI is too basic for real work”).
Ask:
“Do any of these feel familiar or match what we’re seeing?”
Use the pattern cards as:
Prompts for team reflection
Labels to help name friction clearly
Mirrors to surface what people might not be saying out loud
Example:
Pattern: “Thinks AI is too basic for real work”
Signal: “They say the tool’s fine, but they keep tweaking everything by hand.”
Log matched patterns in your Drift Signals Grid.
Step 4: Plot drift signals by impact level
Use the Drift Signals Grid to sort by severity:
(Optional) Step 5: Choose simple changes to reinforce the behaviour
Ask:
“What’s one thing we can do to get this behaviour back on track?”
“Who owns the lever that needs adjusting?”
Fix types:
Prompt tweak (clearer nudge, better UI wording)
Ritual reminder (mention in stand-up, checklist step)
Visibility boost (track progress, shout-out success)
Incentive nudge (recognise correct use publicly)
Clarify roles or instructions (who owns what, when?)
Log chosen actions to track. Assign an owner and a check-in point.
Resources
Task Target Sheet
Define the behaviour you’re scanning for.
Drift Types
Use these different drift types to pinpoint what's going on.
Drift Signals Grid
Plot drift by impact level:
Other methods within the adapt block
Human-AI Performance
© 2026 Alterkind Ltd. All rights reserved.
Human-AI Performance™ is a proprietary methodology developed by BehaviourStudio 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@laurenakelly.com to discuss licensing terms.
Icons by Creative Mahira, The Noun Project.
Thanks to Nicholas Edell, Valentina Tan and multiple VPs implementing AI for your feedback during development.
LICENSE
Based on work by Lauren A Kelly.
For commercial licensing contact: lauren@laurenakelly.com