DocsOpen Henkan Research
Interviews and results

Synthesis

Generate a synthesis of completed interviews, filter it, correct it and validate it before sharing.

Up to date with v1.2.1Updated 20 October 20262 min read

The synthesis turns the completed interviews into findings for each of your objectives, a verdict on each hypothesis, new ideas, reactions to visuals and A/B test ideas. Every statement is backed by quotes linked to the exact moment in the interview.

The validated synthesis of the demo study, shared with the client.
The validated synthesis of the demo study, shared with the client.
1
Synthesis. Results has three views: Synthesis, Transcripts and Ask your interviews.
2
Filter. Choose which interviews to include.
3
Regenerate. Writes a new draft from the interviews.
4
Status. Draft or Validated, and how many interviews it is based on.
5
Client report. The read-only link to share once validated.

Generate the synthesis#

  1. Open Results In the study, select the Results tab. It is available as soon as one interview is completed.
  2. Generate Select Generate the synthesis. The AI reads each transcript, groups findings by objective, checks the hypotheses and selects quotes.
  3. Review The synthesis is a Draft. Read it and correct what needs correcting.
  4. Regenerate after new interviews A dot on the Results tab means interviews finished since the last synthesis. Select Regenerate the synthesis; the new draft replaces the previous one and must be validated again.

Tests are never included. With few interviews, the synthesis gives counts ("3 of 5 participants") rather than percentages.

What the synthesis contains#

SectionContent
SummaryThree to five sentences: the key takeaways and what to do.
Findings by objectiveOne or two findings per objective, each with a title stating the insight, a detail, how many participants are concerned ("5 of 7 participants") and two or three quotes.
HypothesesEach hypothesis with a verdict (Confirmed, Refuted or Unclear), the evidence and quotes.
New ideasTopics participants raised spontaneously, outside the prepared questions.
Reactions to visualsHow participants reacted to each visual shown.
A/B test ideasTwo to four test ideas, each with a hypothesis, a concrete change and a success metric.
Findings by objective, with linked quotes.
Findings by objective, with linked quotes.
1
Objective. The objective this finding answers, copied from your need.
2
Participants concerned. Out of the interviews included.
3
Quote. Opens the interview at this moment.

Quotes stay in the participant's language. Select a quote to open the interview, scrolled to that moment.

Verdicts on the study's hypotheses, then new ideas.
Verdicts on the study's hypotheses, then new ideas.
1
Hypothesis. Copied from your need.
2
Verdict. Confirmed, Refuted or Unclear.
3
New ideas. Topics raised spontaneously.

Filter the interviews#

Select Filter to build the synthesis from a subset of interviews, then generate or regenerate it.

The Filter interviews dialog.
The Filter interviews dialog.
1
Profile. Participants' profile, such as Abandoned cart.
2
Source. Email, public-link tag or on-site.
3
Profile question. Filter by an answer to a profile question.
FilterOptions
ProfileProfiles of participants with a completed interview
SourceThe sources of completed interviews (email, public-link tags, onsite)
Site pagePages where on-site participants were recruited (shown when there are any)
LanguageEN or FR
Profile question and AnswerOne of your profile questions and one of its answers

A filtered synthesis says "(filtered)" next to the number of interviews.

Edit and validate#

  1. Edit Select Edit. Every text becomes editable, and verdicts become a list. Changes save automatically.
  2. Finish Select Finish editing, or Cancel to drop your changes.
  3. Validate Select Validate the synthesis. Its status becomes Validated.

Editing a validated synthesis moves it back to Draft. Only a validated synthesis can be shared: see Share the client report.

For a client organisation, validating also records the study as completed in the client's Henkan account (once per study, even if you validate it again after edits). Studies of the Henkan & Partners organisation are not recorded.