SousMeow
Research & Insights Intermediate Available now Free

Synthesize Interview Notes

Turn interview notes into themes you can act on.

Raw interview notes can turn into confident-sounding fiction if you summarize too quickly. This Cookbook helps you frame the synthesis, extract only supported facts, cluster themes, separate fact from inference, and package a short memo. Enter the real notes and repeated observations you already have. Every prompt forbids invented quotes, demographics, and findings not present in the notes or approved artifacts.

Made for
Researchers, founders, product teams, and community organizers turning interview notes into careful decisions
You leave with
synthesis frame, facts-only extract, theme clusters, fact/inference split, implications, and a research memo
Time
about 50 minutes
Steps
6 guided steps
Requires
any AI you already use (a free chat account is usually enough); sample answers included where available
Completed runs
129 (demo metric)
Average rating
4.7 / 5 (demo metric)

Also in Start Here Selected by SousMeow Recently Added Deep Workflows

Start free to begin

Have an account? Sign in.

Stages

4 stages in sequence
  1. 1

    Frame

    Name the decision and the evidence boundary before summarizing.

    1 step
  2. 2

    Evidence

    Extract facts from notes and cluster only what is supported.

    2 steps
  3. 3

    Themes

    Separate facts from inferences and write careful implications.

    2 steps
  4. 4

    Act

    Package a memo with limits and next moves.

    1 step

Workflow steps

6 steps · completed in order

Stage 1: Frame

  1. 1

    Frame the synthesis goal

    Name the question, decision, evidence boundary, and deadline before touching themes.

    A synthesis without a frame can become a pile of attractive claims. This step says what the notes can and cannot decide. Common mistakes: turning five conversations into a market claim, inventing participant demographics, or skipping the decision context. Need help if the frame sounds too broad? Paste again and ask for a narrower decision. You are ready when the boundary is explicit.

    • Decision is named The synthesis has a concrete decision to inform.
    • Evidence boundary is honest The frame limits claims to the notes and interview count.
    • No invented participant facts No demographics or traits appear unless they were in the notes.

Stage 2: Evidence

  1. 2

    Extract facts only

    Pull supported observations from the raw notes before naming themes.

    Themes should grow from evidence, not from what you hoped to hear. This step keeps raw observations separate from interpretation. Common mistakes: inventing quotes, rounding vague notes into exact counts, or adding demographics from memory. Need help if a bullet sounds like a conclusion? Paste again and ask for facts only. You are ready when each item can be traced to notes or pantry.

    • Facts trace to notes Each observation can be found in raw notes or pantry fields.
    • No invented quotes or demographics The extract does not create quote text or participant traits.
    • Gaps are explicit The output names what the notes do not prove.
  2. 3

    Cluster into themes

    Group supported observations into careful themes and keep one-offs separate.

    Clustering turns evidence into usable structure, but it can also hide weak support. This step asks for themes, supporting notes, singletons, and a reminder that clusters are not universal findings. Common mistakes: naming a theme from one comment, inventing quotes, or treating interviews as a survey. Need help if themes sound too polished? Paste again and ask for plainer cluster names. You are ready when every theme has support.

    • Themes have support Each cluster points back to supported notes.
    • One-offs are separated Singletons do not get promoted into findings.
    • Claims stay proportional The output avoids representative or survey-like claims.

Stage 3: Themes

  1. 4

    Separate fact from inference

    Label what the notes show, what you infer, and what you still cannot claim.

    Teams often act on inferences while calling them facts. This step makes the distinction visible so decisions are safer. Common mistakes: inventing participant profiles, turning concern into demand, or hiding low confidence. Need help if inference language sounds certain? Paste again and require "may" or "could" where support is thin. You are ready when facts and interpretations are separate.

    • Facts and inferences are separate Direct evidence and interpretation appear in different sections.
    • Confidence is qualified The notes explain where support is stronger or weaker.
    • Unsupported claims are blocked The do-not-claim list includes invented quotes, demographics, and findings.
  2. 5

    Write implications

    Turn themes into careful action implications without pretending they are final proof.

    Research becomes useful when it informs a choice, but implications must stay tied to evidence. This step says what to consider next and what not to decide yet. Common mistakes: treating themes as requirements, inventing participant quotes, or deciding strategy from thin notes. Need help if an implication jumps too far? Paste again and ask for a smaller next move. You are ready when every implication has a cautious evidence trail.

    • Implications trace to themes Action ideas come from approved evidence, not new claims.
    • Open questions remain open The output does not pretend unanswered questions are settled.
    • Overreach is named Risks include overreading or ignoring evidence.

Stage 4: Act

  1. 6

    Pack the research memo

    Assemble a concise memo with themes, evidence, next moves, and limits.

    A memo should help people act while remembering what the research can and cannot prove. This final step keeps the summary short and caveated. Common mistakes: adding invented quotes for color, hiding sample size, or writing findings as universal truth. Need help if the memo sounds too confident? Paste again and ask for stronger limits. You are ready when a decision-maker can see both the signal and the limits.

    • Memo stays within evidence The summary does not claim more than notes and approved artifacts support.
    • No invented quotes Themes use paraphrased evidence or exact note text only.
    • Next moves are actionable Recommendations can guide the decision without pretending unknowns are solved.

Your information

7 details, answered once
  • Research question What you hoped to learn from the interviews. Keep it close to the real study.
  • Number of interviews Use the real count. This helps keep claims proportional.
  • Raw notes dump Paste notes, bullets, transcript excerpts, or facilitator notes. Do not clean them up too much.
  • What you heard repeatedly Patterns you noticed before synthesis. Use your own words and do not add new facts.
  • Surprises Anything that challenged your assumption. Leave blank if nothing surprised you.
  • Decision this informs The product, service, policy, or research decision that will use the synthesis.
  • Deadline When the memo or decision is needed.

Finished files

What you export when every step is approved

One Markdown file per approved step, plus kit.html (opens in any browser, offline) and a README of what you approved.

  • kit.html Offline HTML reader for every approved step
  • 01-frame-synthesis-goal.md Frame the synthesis goal
  • 02-extract-facts-only.md Extract facts only
  • 03-cluster-into-themes.md Cluster into themes
  • 04-separate-fact-from-inference.md Separate fact from inference
  • 05-write-implications.md Write implications
  • 06-pack-research-memo.md Pack the research memo
  • README.md Manifest of what you approved
Nearby guided projects you can start next
Research & Insights Available Compare Three Competitors Compare three competitors and leave with a clear next move. 8 steps · 55 min · Intermediate · 4.6 ★ · Free