Synthetic cohort

A Synthetic cohort is built from your community/survey data that allows you to generate insights without launching a new study.

It acts as a research workspace where:

• Your selected data source is organized into a cohort
• AI synthesizes past responses
• You can ask a set of questions to generate insights

Follow these steps:

1. Go to the Synthetic platform
2. Select data source: Choose the community (or dataset) you want to use.
3. Click “New research”

4. Define your cohort:
   1. Name:- Provide a name to your cohort
   2. Short description: Add a brief summary.
   3. Cohort purpose: Provide a detailed cohort purpose: The system will use this purpose to filter the data from the community and create a cohort to run different research modules like IDI and Converse AI.

      Example:

      • Understand why customers churn
      • Explore perceptions about pricing
      • Evaluate onboarding experience
   4. Set a similarity cut-off percentage: Choose the similarity percentage that determines how closely members should match your cohort purpose.
      • Higher % (e.g., 80–90%): More precise match, smaller but highly relevant cohort
      • Lower % (e.g., 50–70%): Broader match, larger cohort with more diversity
   5. Save and generate: Click Create

Your Synthetic cohort is now ready to use for insight generation.

Post-creation, access two main features: In-Depth Interviews (IDI) and Converse AI.

1. In-Depth Interviews (IDI)
   • Select cohort members (1-50) based on matching scores to your study purpose.
   • Add or AI-generate questions.
   • Run interviews → Get AI responses, key takeaways, summaries, quotes, similarities/differences, and action recommendations.
   • Download report as PDF or Markdown.
2. Converse AI
   • Engage in real-time, conversational interviews with cohort members.
   • Ask dynamic follow-up questions based on responses.
   • Ideal for exploratory research, probing deeper insights conversationally.
   • Generate summaries or export chat transcripts.

Both preserve privacy—no real data is used beyond synthesis, and results stay in your account.