PM Mapped
Working with AI SYNTHESIS · Affinity Mapping

Affinity Mapping with AI

Affinity mapping — clustering many observations into themes — is exactly the kind of synthesis AI can accelerate. But the point of the exercise is the thinking that happens while you cluster: the felt sense of a pattern forming. Let AI seed and challenge your map, then do the sense-making yourself.

← Back to the Affinity Mapping tool
1How to use AI for this

A step-by-step way to work

1

Get a first-pass layout

Paste your observations and ask AI to propose affinity groups with candidate theme labels — a starting layout, not a verdict.

2

Hunt the outliers

Ask what doesn't fit any cluster. Outliers are often where the interesting insight hides.

3

Re-cluster deliberately

Ask it to regroup the same notes a completely different way to break your own framing and reveal alternative structures.

4

Pull the layout into your own space

Move the clusters into your tool and physically rearrange them — the manual sense-making is the part that produces insight.

5

Name themes yourself

AI's labels are generic ('Usability issues'); the sharp, insight-bearing label is yours to write.

2Worked examples

Real prompts, and what good output looks like

First-pass clusters and outliers
Your prompt

Cluster these 15 research observations into affinity groups, label each, and separately list anything that doesn't fit a group [observations pasted].

What good output looks like

Group 1 – Setup friction (5 notes) Group 2 – Trust/credibility concerns (4) Group 3 – Mobile gaps (3)

Outliers: 'used it differently than intended — to share with their accountant' (1); 'asked if data could be deleted on request' (1).

The first outlier hints at an unanticipated use case; worth exploring before you dismiss it.

The outliers are the gold: AI surfacing the 'shared with accountant' note may point to a use case you didn't design for. Don't let neat clusters bury it.
Forcing a re-cluster
Your prompt

Now regroup the same 15 observations by the user's emotional state instead of by feature area.

What good output looks like

Anxious / uncertain (6): mostly setup and trust notes — these cluster around the first-time experience.

Confident / in-flow (4): power-user behaviours.

Frustrated (5): spread across features — suggesting frustration isn't feature-specific but moment-specific.

This cut suggests intervening by moment (first-run) may beat fixing features one by one.

Why re-cluster: the emotional cut revealed a 'fix the first-run moment' insight the feature-cut hid. AI is great for breaking your default framing.
3Copy-paste template

A prompt you can reuse

Fill in the highlighted parts and paste it into your AI tool of choice. Edit the output — it's a starting point, not a finished answer.

Reusable prompt
Help me affinity-map research observations. Give me a starting layout to think with — I'll do the final sense-making.

Observations:
[paste your observations, one per line]

Please:
1. Propose affinity groups with a candidate label for each and the count per group.
2. Separately, list any observations that don't fit a group (outliers) and why they're interesting.
3. Then re-cluster the SAME observations a different way (e.g. by emotional state, by journey stage) and tell me what that alternative cut reveals.
4Common pitfalls

What AI gets wrong here

Skipping the thinking

Accepting AI's finished clusters skips the synthesis that actually generates insight.

Do this instead: Use the layout as a seed, then rearrange it yourself — the manual act is where patterns click.

Burying outliers

Tidy clusters can absorb or hide the odd note that's the real discovery.

Do this instead: Always ask for outliers explicitly and treat them as leads, not noise.

Generic theme labels

AI labels themes blandly ('Usability'), which carry no insight.

Do this instead: Rewrite labels yourself into specific, surprising statements (this connects to the Insight Statement guide).
The judgment that stays yours

AI can group sticky notes faster than you can, and forcing it to re-cluster is a genuine way to escape your own framing. But affinity mapping earns its keep through the act of clustering — the moment a pattern becomes felt. Outsource the first draft if you like; do the sense-making, the outlier-chasing, and the naming yourself.