Quick-Start Guide: Prompting Interviews and Focus Groups

We know that analyzing long, complex interview and focus group transcripts can be challenging, even with AI. The volume of back-and-forth dialogue often makes it hard to get focused, high-quality results. The fastest way to get focused, trustworthy insights from multi-speaker data in the AI View is to use prompts that leverage the Speaker ID and timeframe features.

Forget broad summaries—start here with powerful examples you can copy and paste to instantly drill down into specific speakers, topics, and moments.

Considerations for Analysing Imported Interviews and Focus Groups

Indeemo supports not just diary-based research, but also the import of video files such as recorded interviews and focus groups. These video files are automatically transcribed. This makes it possible to apply AI analysis at scale, but there are important differences in how you work with this data compared to diary studies.

Speaker Role Assignment Is Essential

To ensure the AI interprets focus group or interview transcripts accurately, speaker roles must be assigned by the user. These roles are simple but powerful:

• Researcher: The moderator, interviewer, or facilitator.

• Respondent(s): The participant(s) being studied.

You can also assign individual names to speakers (e.g., "Stephen," "Hannah," "Tom"), which opens up the possibility of prompting the AI to compare or focus on specific individuals.

Pro Tip: Your AI analysis is only as good as your Speaker Role Assignment. Before you run any analysis, you must assign speakers as either Researcher or Respondent(s). This is an important setup step! It ensures the AI ignores the moderator's comments and only processes the content tagged as researcher or respondent, significantly improving analysis quality.

Identifying Topics Covered by the Researcher

Use this for quickly building an index or table of contents for an interview or focus group, allowing you to see the main discussion topics and where they occurred. This is a crucial pre-step for creating focused follow-up prompts.

💡 The Breakdown: Why This Prompt Works

• Context/Task: The prompt clearly defines the job: "Identify a list of topics discussed". It narrows the data source by explicitly asking for topics "as outlined by the researcher," using the Researcher Speaker ID to structure the output.

• Focus Area/Details: It requires a timestamp for each topic, which helps you quickly locate discussion boundaries and then formulate the time-specific prompts (like the one in section 3).

• Constraints: Requiring a "brief description" ensures the AI organizes the raw topic title into a usable overview.

• Structure: The structure is implied as a clear, easy-to-scan list of topics with associated details.

Summarizing an Individual's Contribution

This prompt is designed to help the AI extract only the key insights, quotes, and perspectives contributed by a single person in a multi-participant setting (like a focus group), which is crucial for building detailed case studies or personas.

💡 The Breakdown: Why This Prompt Works

• Context/Task: The instruction is highly specific: "Summarise the key insights" and the use of the speaker's name tells the AI to isolate that speaker's dialogue from the rest of the group.

• Focus Area: By asking for a summary of "key insights," you steer the AI away from factual reporting and toward high-level findings.

• Constraints: It forces the AI to provide evidence by requiring quotes and a timestamp for each insight, making the individual's profile easy to verify and present.

• Structure: The output will be an organized list of high-level insights, each supported by direct evidence from the transcript.

Analyzing a Specific Timeframe and Topic

Use this prompt to drill down into a short, critical section of a single, long transcript. This is ideal when you need to focus on a particular discussion point without analyzing the entire hour-long interview.

💡 The Breakdown: Why This Prompt Works

• Context/Task: You define the precise timeframe and the task ("what recurring themes did respondents discuss"), allowing the AI to skip irrelevant portions of the transcript.

• Focus Area: Specifying "respondents" ensures the analysis is participant-driven, ignoring any researcher/moderator dialogue within that time window.

• Constraints: It limits the output to recurring themes and requires a quote with a timestamp as supporting evidence, ensuring you can verify the finding against the source video.

• Structure: The structure is clear—a list of themes, each with an example, ensuring clean, actionable output.

Comparing and Contrasting Individual Speakers

This advanced use case helps you segment your audience and build detailed personas by analyzing and comparing specific participants within a focus group or across multiple interviews.

💡 The Breakdown: Why This Prompt Works

• Context/Task: It clearly states the goal: a comparative analysis. By providing the names (e.g., Shane and Stephen), you leverage the assigned Speaker IDs to focus the analysis.

• Focus Area: It narrows the lens by specifying the topic ("experience using a food delivery app"), ensuring the AI ignores unrelated dialogue from these speakers.

• Constraints: It requires a structured output ("organized report with sections") and forces the AI to provide evidence by requiring quotes and a timestamp for each finding, making the insights trustworthy.

• Structure: The structure is implied (a side-by-side or thematic comparison), ensuring clean output for immediate reporting.