NotebookLM (Now Gemini Notebook): Using Google’s AI Notebook in Business

What Is NotebookLM?

NotebookLM is an AI tool from Google that works differently from ChatGPT or Gemini in chat: it doesn’t answer from a language model’s general knowledge, but exclusively from the sources you load into a notebook. You ask questions of your own documents, get answers with citations, and can turn the material into summaries, outlines, FAQs, flashcards or a podcast-style audio summary. Under the hood, it’s a fully built RAG system with an interface anyone can use.

Since 16 July 2026 the product has been called Gemini Notebook. Google has folded it into the Gemini family; existing notebooks and links keep working. According to Google, more than 30 million people and over 600,000 organisations use the tool. In everyday usage — and in search queries — “NotebookLM” remains the familiar name for now; this article uses both.

How Does Gemini Notebook Work?

The workflow comes down to three steps: create a notebook, upload sources, ask questions. As sources, the tool accepts PDFs, text files, Markdown, Google Docs and Slides, websites via URL, YouTube videos (via transcript), audio files, and pasted text. Each source is analysed and turned into an automatic summary. The chat then answers questions across all sources — with numbered citations that jump to the original passage on click.

From the sources, you can generate content with a single click: a study guide, a briefing document, a timeline, a mind map, flashcards, a quiz, an infographic sketch, a presentation, or an audio summary in which two AI voices discuss the content — available in German as well since early 2026. New since July 2026 is a secured cloud computer per notebook that can write and run code, for example to analyse tables drawn from the sources. The feature is initially available to AI Ultra and Workspace customers and is being rolled out to Pro users in stages.

What’s the Difference Between NotebookLM and ChatGPT?

The tools complement each other: NotebookLM is the reading and research tool, ChatGPT or Gemini the writing tool. Anyone deploying both in their business should know the framework we set out in ChatGPT for Business.

Can You Use NotebookLM for Free — and What Does More Cost?

Yes. With a personal Google account, Gemini Notebook is free. The limits are generous for getting started:

Seven Business Use Cases

• ▸Onboarding: load the handbook, process descriptions and policies into a notebook — new hires ask instead of searching, and listen to the audio summary on their commute.
• ▸Tenders and contracts: upload lengthy documents and ask targeted questions about deadlines, obligations and exclusion criteria — with a source reference to check.
• ▸Meetings: collect transcripts or minutes, and query decisions and open points across weeks.
• ▸Market research: condense studies, competitor websites and industry reports into one notebook; use the briefing document as a starting point for strategy.
• ▸Product knowledge in sales: data sheets, price lists and FAQs — the team gets sourced answers in seconds for customer conversations.
• ▸Training and exam prep: flashcards and quizzes generated from training materials, for example on workplace safety or compliance.
• ▸Content groundwork: summarise and structure your own sources before writing copy — so the content stays grounded rather than invented.

For recurring, team-wide knowledge queries with permission management and integration into a chatbot or intranet, a single notebook eventually hits its limits. At that point, a dedicated RAG system is the next step — NotebookLM is a good way of first establishing which questions the team actually asks.

Is NotebookLM GDPR-Compliant?

The answer depends on the account. With personal Google accounts, uploaded content is, according to Google, not used for training unless you actively give feedback — in which case reviewers can see the data. There is no data processing agreement (DPA), and the data sits on US infrastructure. A personal account is therefore unsuitable for trade secrets or personal data.

With a Google Workspace account, the Workspace data protection terms apply: uploads, queries and answers are not manually reviewed and are not used to train AI models, even with feedback; Workspace’s data processing agreement applies. Google does not currently offer EU-only hosting specifically for Gemini Notebook — anyone who needs exclusively European processing is better served by a self-hosted RAG system or a local LLM.

• ▸Use it only via Workspace accounts, never via personal accounts.
• ▸Don’t upload health, salary or other especially sensitive data unless strictly necessary.
• ▸Share notebooks via permissions rather than downloads; revoke access when employees leave.
• ▸Add the use case to your AI policy and document it in your record of processing activities.

Limitations and Common Mistakes

• ▸Source quality: outdated or contradictory documents lead to contradictory answers. Clean up before uploading.
• ▸Scanned PDFs: image-based documents without a text layer are poorly recognised — run OCR first.
• ▸No substitute for reading: the source citations are the big advantage — use them. A summary is a condensation, not a verification.
• ▸Audio summaries: entertaining and good for onboarding, but not built for precision; the hosts simplify.
• ▸Daily limits: the free tier’s 50 chat queries are quickly used up when a whole team is on it.

Our Verdict

Gemini Notebook is the most accessible tool for putting AI to work on your own documents — free, usable immediately, and with source citations that ChatGPT doesn’t offer in this form. For individuals and small teams, it’s a genuine productivity gain for research, onboarding and document work. Businesses should run it consistently through Workspace, and know where the boundary with a dedicated knowledge system lies.