The MindMap Standard
AI policy
Artificial intelligence speeds up certain tasks at MindMap. It decides none of them, and it never speaks in place of a respondent.
What AI never does
- Generate a synthetic respondent to fill a quota, even partially.
- Extrapolate a missing human response from a statistical or generative model.
- Produce a result, conclusion or recommendation without an analyst's review.
- Replace a qualitative interview with a simulated conversation.
What AI does, under supervision
- Transcription
- Recorded qualitative interviews are AI-assisted in transcription, then reviewed in full by the analyst running the study.
- Verbatim coding
- A model proposes a first thematic coding pass on open-ended responses; a sample is manually audited before the code plan is validated.
- Translation
- First drafts of translation (questionnaire, discussion guide) are AI-assisted, then back-translated and validated by a human linguist.
- Internal formatting
- Report templates, meeting summaries and internal materials may be AI-assisted. Never the findings or analysis delivered to the client.