Help Us Notice Nuance · Using AI for Research

You have a body of material. Now the relationship changes. AI helps surface patterns, connections, tensions, and missing perspectives — and can turn half-formed participant input into something you can talk about.

Sharper conversation — not a flatter summary.

This is not “summarize the folder.” Summaries flatten. Nuance work should make the conversation sharper.

Watch for clusters with no quotes, themes that only echo your prompt, and prototypes that persuade you more than they reveal the participant.

Core principles

Candidate themes (max 5) each need supporting quotes or paraphrases tied to a source.
Preserve tensions — disagreement and self-contradiction are load-bearing.
Name missing perspectives and weak signals; do not resolve them yet.
Generated artifacts are probes (“what feels right / off?”) — mismatch often reveals what mattered.
Human checkpoint: re-read one primary source; name one surprise; name one claim you will not stand behind yet.

Check yourself

What marks good AI-assisted synthesis in Notice nuance mode?

Do this in Copilot

Run tension-keeping synthesis on your narrowed set. Save themes and tensions — Lesson 4 attacks the themes on purpose.

Paste this into Copilot Chat and work through it before moving on.

Synthesis that keeps tension

I will paste research notes / transcript excerpts about [TOPIC]. Produce:
1. Candidate themes (max 5) — each with 2 short supporting quotes or paraphrases tied to a source
2. Tensions — where people disagree or contradict themselves
3. Missing perspectives — who is not in this material but should be
4. Weak signals — odd or rare comments that might matter
5. What you might be projecting — guesses, not evidence
Do not write recommendations yet. Do not resolve the tensions.

Techniques: Structured synthesis · Counterexamples

If every theme feels tidy, ask: “Give the strongest counterexample to each theme.”

Open Copilot

Did you run this in Copilot? Mark complete when you have tried it.