Define the change before you plan it. The single biggest driver of resistance is ambiguity — start here so everything downstream has something solid to point back to.
What kind of change is this?
Project — has a defined end date
Ongoing — no fixed notice or end point
Delivery approach
Big bang
Phased
Pilot & scale
Continuous
Hybrid
Who's affected — internal and external. Every change touches people beyond the immediate team. Map both sides here; missing an external stakeholder is usually where surprises come from.
Stakeholder / group
Internal / External
Impact
Interest
Key concern
Engagement plan
Kotter's 8-step model. The first four steps unfreeze the status quo; the next three drive movement; the last anchors the new way. Track status and add your own notes at each step — most failed change efforts skip straight to step 5 or 6.
The Change Curve. People move through this emotionally, not logically — often at different speeds, and not always in a straight line. Productivity typically dips before it recovers. Support the stage people are actually in, not the stage you wish they were in.
1 · Denial / Shock
"This won't affect us" or "just another initiative." Respond with: over-communicate the facts and the vision, and listen without arguing.
2 · Resistance / Anger
Fear, pushback, lowest productivity. Respond with: empathy, involve people in shaping the solution, address "what's in it for me."
3 · Exploration
Curiosity returns; people start testing the new way. Respond with: coaching, safe spaces to try things, quick feedback loops.
4 · Commitment
New behaviour is embedded. Respond with: celebrate it, reinforce it in how you review and reward work.
Who's where, right now
Track each stakeholder group's current phase — this will move, and that's expected.
Stakeholder group
Current phase
What support they need right now
Two different risks. A risk register covers what could go wrong. Regression tracking covers something easier to miss — people quietly reverting to the old way once attention moves elsewhere.
Risk register
Every live risk, with likelihood, impact and what you're doing about it.
Risk
Likelihood
Impact
Mitigation
Regression tracker
What old habit could resurface, the early warning sign, and how you'll actually notice.
Old behaviour that could return
Early warning sign
How we'll monitor it
Dependencies
What this change relies on outside your direct control — supply chain, another team, a system, a third party.
Dependency
Type
If it's disrupted
Mitigation
What it costs, and when it happens. Cost implications are rarely just budget — factor in time, training and the disruption cost of doing nothing.
Cost implications
Item
Estimated
Actual
Notes
Timeline & milestones
Milestone
Target date
Owner
Status
Notes
Readiness score — before you launch. Score each area honestly: Red (not ready), Amber (partly there), Green (ready). This isn't a box-tick — a low score somewhere real is more useful than a comfortable average.
Put AI to work on your change
Leading change means holding a lot at once — stakeholders, risks, resistance, timing. A capable AI can help you think through all of it faster, if you direct it well. These prompts use the 4Ds from SMART AI: a clear role, your real context, a constraint, and a request to show its reasoning so you can check it.
Replace the [bracketed] parts with your own detail and paste into whichever AI you use. The italic note under each explains why it's built that way.
Map the stakeholders you're missing
Description
Use on the Stakeholders tab when you've listed the obvious people but suspect there's a blind spot.
You are a change-management adviser. I'm planning this change: "[describe the change, who it affects, and the stakeholders I've already identified]". Who am I likely to be missing — the quiet influencers, the people affected downstream, the ones who can block without formal power? For each, note their probable concern and how much they matter. Show your reasoning, and flag which one I'd most regret overlooking.
Why this shape: naming the categories you tend to miss (downstream, informal blockers) prompts the AI past the obvious org chart — where most change plans have their gaps.
Surface the risks you're not naming
Discernment
Use on the Risk & Regression tab to pressure-test your risk register before it goes to anyone senior.
Act as a pre-mortem facilitator. Assume my change has failed six months from now. Here's the plan: "[paste your change, timeline and the risks you've logged]". Working backwards from failure, give me the three most likely causes I haven't already listed — including the human ones, like people quietly reverting. For each, an early warning sign I could watch for. Be blunt; I'd rather hear it now.
Why this shape: the pre-mortem ("assume it failed") is a proven technique that unlocks risks a normal "what could go wrong?" never surfaces — and it works especially well with AI.
Anticipate objections before the room does
Delegation
Use before announcing a change, to prepare honest answers to the pushback you'll actually get.
You are a sceptical but fair member of my team. I'm about to announce this change: "[describe the change and why it's happening]". Give me the five hardest, most realistic objections you'd raise — the ones said in the corridor, not the meeting. For each, don't just rebut it; tell me what legitimate concern sits underneath it that I should actually address. Keep each to a couple of lines.
Why this shape: "said in the corridor, not the meeting" gets you the real objections, and asking for the legitimate concern underneath keeps you from just arming yourself with rebuttals.
Write the change message three ways
Diligence
Use when you need to communicate the change to different audiences who'll each hear it differently.
Act as a clear, humane internal-comms writer. Here's a change I need to communicate: "[describe the change, the reason, and what's genuinely uncertain]". Draft three short versions of the message — one for the people most affected, one for their managers, and one for the wider organisation — each honest about what's known and unknown. Don't spin it. After each, note the one line most likely to be misread, so I can check it.
Why this shape: "don't spin it" and flagging the misreadable line push against AI's tendency toward corporate gloss — and remind you to check tone before sending.
One rule that keeps you safe. AI is a fast thinking partner, but it doesn't know your people and it will sound certain when it's guessing. Treat every answer as a well-read colleague's opinion: useful, worth testing, never the final word — especially on anything involving named individuals.