Findings and Insights Summary

What this does

Turns a pile of raw discovery material into themed findings you can stand behind. What people said and did (evidence) sits clearly apart from what it means (interpretation), so when a client challenges a conclusion, you can point straight at the material behind it. The gap between what people say they do and what actually happens - usually the most valuable thing discovery surfaces - gets its own spotlight.

What you give it

• The raw material - interview notes, workshop outputs, observations, document reviews, survey comments, pasted or attached
• The questions the discovery was meant to answer, in a line or two (optional - themes will be derived from the material if unstated)
• How sources should be referenced - by role, by number ("Interview 3"), or anonymised entirely (optional - roles are used by default)
• Anything already known that the findings should test rather than repeat (optional)

How it does the work

1. Read the entire body of material before forming a single theme. Themes built from the first two interviews get quietly contradicted by the sixth.
2. Ask up to 3 short questions if essentials are missing - what discovery was meant to find out, how to cite sources, or whether one loud voice in the material should be weighted or flagged. Then proceed.
3. Tag the material as you go: statements (what people said), observations (what was seen or documented), and artefacts (what the documents show). Keep the source attached to every item - an unattributed finding can't be defended later.
4. Group the tagged material into 3-6 themes. A theme earns its place through weight of evidence - multiple sources, or one source with strong corroborating documents. A single interesting comment isn't a theme; park it in "worth noting".
5. For each theme, write the evidence first: what was said and done, with short quotes and source citations. Quotes are verbatim or they aren't quotes - trim with ellipses, never improve the wording.
6. Then, separately, write the interpretation: what this evidence most likely means, and how confident the material lets you be. Label it as interpretation. Where two readings are plausible, give both.
7. Hunt the say-do gaps deliberately: places where what people said (or what the official documents claim) diverges from what the observations and artefacts show. Write each as its own finding - stated position, observed reality, evidence for both sides.
8. Record what's missing: voices not heard, questions unanswered, themes resting on thin evidence. Mark these [to confirm] - they shape the next round of discovery and protect the user from overclaiming.
9. Draft using the output layout. Order themes by strength of evidence, not by interest.
10. Do a final pass: every interpretation traceable to cited evidence, no recommendations creeping in, nothing that reads as blame. Show the draft and name the two or three findings most likely to be challenged.

Output layout

# Findings and Insights - [Engagement name]

**Based on:** [sources - e.g. 8 interviews, 2 workshops, document review]
**Date:** [date]

## Overview
[Three or four sentences: what discovery set out to learn and the shape of what it found. No conclusions yet.]

## Theme 1 - [Name, in plain words]

### What we heard and saw
- "[Short verbatim quote]" - [source]
- [Observation or document evidence] - [source]
- [Evidence item] - [source]

### What we think it means
[Interpretation, clearly labelled as such. Confidence level in plain words - e.g. "consistent across six of eight interviews" or "based on two sources, so hold it lightly".]

[Repeat per theme.]

## Where words and reality part ways
| What's said | What we observed | Evidence |
|-------------|------------------|----------|
| [Stated intent or official position] | [What the material shows happening] | [Sources for both] |

## Worth noting
- [Single-source item that didn't make a theme but shouldn't be lost] - [source]

## What we don't know yet
- [Gap - voice not heard, question unanswered] [to confirm]
- [Theme resting on thin evidence] [to confirm]

Rules

• Australian English with contractions - organise, prioritise, we'll, they've.
• Evidence and interpretation never share a sentence. The reader must always know which one they're reading.
• Never invent quotes, counts or sources. Quotes are verbatim; "several people said" means you can count them; gaps are [to confirm].
• Findings describe, they don't prescribe. Recommendations belong in the playback, not here.
• Contradictory evidence goes in, not out. A theme that ignores its counter-evidence will be dismantled in the playback meeting.
• Cite sources the way the user chose, consistently, and never in a way that exposes someone who spoke in confidence.
• This skill makes sense of the material. Turning it into a client-facing story is the Findings Playback Report skill's job.