How to use NotebookLM, now called Gemini Notebook

What is NotebookLM and what is it good for at work?

NotebookLM, now Gemini Notebook, is an AI research assistant that works from the sources you give it rather than the open web. Google's help centre says you can add PDFs, websites, YouTube videos, audio files, Google Docs and Google Slides, and chat answers come with inline citations back to those sources. Google's Workspace product page adds text and Markdown files and pasted text, with each source holding up to 500,000 words or 200MB for uploads.

From the same sources it can produce study guides, briefings, Audio Overviews, mind maps, flashcards, quizzes and more. Google's July 2026 announcement says the rename keeps it the same standalone product, now also reachable inside the Gemini app, with notebooks syncing between the two.

At work it suits jobs where the answer must come from a fixed set of documents: understanding a policy pack, preparing for an audit, onboarding onto a project's history or getting through a long report before a meeting.

How do I get started with NotebookLM?

Sign in with a personal Google Account or a work or school account; Google says work and school access must be turned on by your administrator, and India is among the supported countries. Open the product from notebook.google or from notebooks in the Gemini app, create a notebook and add your first sources.

Start small: three to five documents on one topic work better than a folder dump. Then ask your first question in chat, something you already half know the answer to, so you can judge the citations: 'What does the leave policy say about carrying over unused days? Quote the relevant lines.' Click each citation and read the passage it points to.

What is a good 10-minute NotebookLM starter exercise?

Use a fictional 25-person charity, Riverside Literacy Trust, onboarding a new volunteer coordinator. Write three short invented documents: a volunteer handbook, a safeguarding summary and last term's programme report.

Minute 1 to 2: create a notebook called Coordinator onboarding and add the three documents. Minute 3 to 5: ask for the ten things a new coordinator must know in their first week, each with a citation. Minute 6 to 7: ask which questions the documents do not answer; a good notebook tells you where the gaps are. Minute 8: pin the best answer as a note, because Google's page says chat history is not currently kept. Minute 9 to 10: generate a briefing or an Audio Overview and compare one claim in it with the cited source.

The skill being practised is source discipline: everything the new coordinator hears should trace back to a document the charity owns.

How do I get better results from NotebookLM?

Because NotebookLM answers from your sources, the biggest improvements come from curating those sources and asking for evidence. These techniques use features Google describes.

What should I not upload to NotebookLM, and what does Google say about data?

Upload only documents your employer's policy allows you to put into this tool, and prefer your work account for work material. Do not add files containing customer personal data or confidential deal terms unless your policy explicitly permits it.

Google's help centre says your data is not used to train Gemini Notebook unless you provide feedback. The Workspace product page says sources stay private unless you choose to share a notebook. For work accounts, Google's Workspace privacy hub states that uploads, queries and responses in Gemini Notebook are not reviewed by human reviewers, and that Workspace customer data is not used to train models without permission.

Where does NotebookLM fall short?

It is only as good as its sources. If the handbook is out of date, NotebookLM will quote the out-of-date rule with a tidy citation. It is also the wrong tool for questions that need the wider web or fresh judgement; that is a job for a general assistant or for a colleague.

Google lists safety flags, unclear phrasing and missing information as reasons it may not answer, and usage limits apply. Generated formats such as Audio Overviews simplify; check anything specific before you rely on it. The rename also means menus and help articles may show either name for a while.

How does Tosheo.ai teach NotebookLM inside a role path?

Tosheo.ai uses NotebookLM for source-grounded tasks in each role: an HR generalist answering policy questions from a handbook, a compliance-minded operations manager preparing for an internal review, or a nonprofit programme lead summarising funder reports. Lessons supply fictional source packs, so you practise without uploading real files.

Each 10 to 15 minute lesson ends with a hands-on task, marked against a visible rubric for citation accuracy, gap-finding and review, with specific feedback. On the Team plan, managers can add company documents to lessons and see skill progress, not chat logs. The 7-day free trial needs no credit card.

Sources and scope

Google Help, Learn about Gemini Notebook, checked 19 October 2026 ↗

Google Blog, NotebookLM is now Gemini Notebook, checked 19 October 2026 ↗

Google Workspace, Gemini Notebook (formerly NotebookLM) product page, checked 19 October 2026 ↗

Google Workspace, Generative AI in Google Workspace Privacy Hub, checked 19 October 2026 ↗