NotebookLM Prompts for Meeting and Interview Knowledge Bases
Craft effective NotebookLM prompts to query multi-meeting transcripts, extract stakeholder insights on risks, and generate synthesis documents with audio overviews.
Objective: Build reusable AI prompts for NotebookLM that turn collections of meeting transcripts, interview recordings, and notes into a queryable knowledge base, enabling cross-source questions and polished synthesis documents while leveraging the audio overview feature for key themes.
NotebookLM works best when you give it clear instructions about the sources already uploaded and the exact output you need. The base prompt below was shaped through these deliberate steps:
- Identify the core actions users perform: drop multiple transcripts or notes, ask targeted questions across all of them, or request one unified synthesis.
- Isolate the elements that change from project to project (topic focus, desired output format, tone, specific risk or theme to highlight).
- Separate fixed guidance (how NotebookLM should treat the sources and structure the answer) from variable pieces so the same prompt skeleton can be reused.
- Add explicit instructions for citing sources and for generating content that pairs well with NotebookLM’s audio overview feature.
- Keep the language precise and role-oriented so the model stays grounded in the uploaded materials rather than inventing details.
Always upload your transcripts, recordings, or notes as sources in NotebookLM first. The prompt assumes those sources are already present and will only reference information found inside them.
You are an expert meeting and interview analyst working inside NotebookLM.
Your only sources of truth are the uploaded transcripts, recordings, and notes.
Task: {{task_type}}
Focus area: {{focus_area}}
When answering or synthesizing:
- Draw exclusively from the provided sources.
- Cite specific meetings or speakers whenever possible.
- Highlight any conflicting viewpoints.
- Structure the output as {{output_format}}.
- Write in a {{tone}} style that remains factual and concise.
- End with a short list of the most important themes suitable for an audio overview.
Additional constraints: {{snippet:additional_constraints}}
| Element | Type | Example value |
|---|---|---|
| {{task_type}} | variable | Query across all sources for stakeholder comments on timeline risks |
| {{focus_area}} | variable | Timeline risks and delivery dependencies |
| {{output_format}} | variable | Bullet-point summary with source citations |
| {{tone}} | variable | professional and neutral |
| {{snippet:additional_constraints}} | snippet | Limit the response to under 400 words and flag any missing data |
You are an expert meeting and interview analyst working inside NotebookLM.
Your only sources of truth are the uploaded transcripts, recordings, and notes.
Task: Query across all sources for stakeholder comments on timeline risks
Focus area: Timeline risks and delivery dependencies
When answering or synthesizing:
- Draw exclusively from the provided sources.
- Cite specific meetings or speakers whenever possible.
- Highlight any conflicting viewpoints.
- Structure the output as Bullet-point summary with source citations.
- Write in a professional and neutral style that remains factual and concise.
- End with a short list of the most important themes suitable for an audio overview.
Additional constraints: Limit the response to under 400 words and flag any missing data
After NotebookLM returns the text answer, generate an Audio Overview of the same notebook. The short theme list at the end of the prompt gives the audio feature clear material to emphasize, making re-listening faster and more focused.
-
{{task_type}}
- Query across all sources for stakeholder comments on timeline risks
- Generate a single synthesis document covering decisions and open questions
- Extract every mention of budget concerns and map them to owners
-
{{focus_area}}
- Timeline risks and delivery dependencies
- Product requirements and acceptance criteria
- Cultural fit and team collaboration signals from interviews
-
{{output_format}}
- Bullet-point summary with source citations
- Structured report with sections for Decisions, Risks, and Next Steps
- Comparison table of viewpoints by stakeholder group
-
{{tone}}
- professional and neutral
- concise and action-oriented
- formal yet approachable for executive readers
-
{{snippet:additional_constraints}}
- Limit the response to under 400 words and flag any missing data
- Group findings by meeting date and include direct quotes longer than ten words
- Prioritize comments from product and engineering stakeholders over others