Introduction
This tutorial fixture demonstrates a source-grounded research workflow. Product controls and availability can change, so replace the interface notes and screenshots after checking the current version.
Requirements and prerequisites
- An account with access to the current NotebookLM interface
- A small set of authoritative sources you are allowed to use
- A clear research question and a place to record verification notes
Step-by-step workflow
1. Prepare the source set
Remove duplicates, check dates and authorship, and begin with the smallest set that can answer the question. Record why each source is trustworthy and relevant.
2. Create a notebook and add sources
Use a descriptive notebook name, add the prepared material, and confirm that each source finished processing. Do not assume a failed or partial import is available to later answers.
3. Ask narrow, checkable questions
Ask one question at a time and request the specific distinction, summary, contradiction, or evidence you need. Preserve the question with the resulting answer.
4. Open every important citation
Check that the cited passage supports the whole statement, not just a nearby phrase. Inspect surrounding context, dates, definitions, and conflicting sources.
Screenshots and examples
Add original screenshots that show source import status, a representative question, the answer, and the cited passage. Captions should explain what each image proves.
Expected result
You should finish with a set of working notes whose important claims can be traced back to inspected source passages. The final deliverable still needs human synthesis and review.
Common mistakes
- Uploading too many overlapping or low-quality sources
- Asking broad questions whose answers are difficult to verify
- Quoting the generated summary without opening the citation
- Ignoring a source’s date, scope, methodology, or conflicting evidence
Limitations
Access, supported source formats, limits, interface labels, model behavior, and regional availability may change. Source grounding helps inspection but does not guarantee complete, unbiased, or correct synthesis.
Conclusion
The reliable habit is simple: curate sources, ask narrow questions, inspect citations, preserve context, and treat the generated answer as an auditable draft rather than a final authority.
