The AI-powered research app is breaking out of its walled garden, pairing its personalized note-taking with a powerful Gemini agent that can autonomously build an entire research report for you.

Google is giving one of its most interesting AI tools a major upgrade. The company announced on Thursday that it’s pairing two of its research products, plugging NotebookLM directly into Gemini’s agentic AI tool, Deep Research. The new capabilities are rolling out to all users and should be available within a week.

This move is a big deal for anyone who uses NotebookLM, transforming it from a “closed-world” assistant into an active research partner that can now browse the open web.

Before this update, you could think of NotebookLM as a smart, private study.

It’s not a blank page like Google Docs. Instead, you feed it your own sources: PDFs of research papers, Google Docs, text you’ve copied, links to websites, and even Google Slides.

Once you upload these sources, NotebookLM uses Gemini to become an expert only on the information you’ve provided. You can then ask it questions, and it will answer based solely on your documents, complete with citations. It can generate summaries, create study guides, and even produce an “Audio Overview,” which turns your notes into a private, on-the-fly podcast.

Until now, its greatest strength was also its main limitation: it only knew what you told it. If the answer wasn’t in your uploaded files, NotebookLM couldn’t help.

That’s all about to change.

The new feature, Deep Research, is what Google calls an “agentic AI tool.”

In plain English, an “agent” is an AI that doesn’t just answer a question; it takes action. You give it a complex goal, and it creates a multi-step plan and executes it on its own. In this case, you’re tasking it with being your personal research assistant.

This integration isn’t just about letting NotebookLM “Google” things. It’s a full-fledged research-and-synthesis engine. As Google recently linked Deep Research to other Workspace products like Gmail and Drive—allowing Gemini to draw context from your emails and files—its integration into NotebookLM is the next logical step.

When you direct a question to Deep Research within NotebookLM, you’ll get two options:

1. Fast Research: This is for a quick lookup. Google says it will “rapidly” look for information. Think of it as a sophisticated, targeted search for fact-checking or grabbing a few quick sources.
2. Deep Research: This is the main event. When you choose this, the AI will perform an “in-depth analysis to find high-quality sources.”

Workflow reimagined: from blank page to knowledge base

This is where the new workflow gets really powerful.

According to Google, when you select the “deep” style, you’ll first get a research plan from the AI, showing you how it plans to tackle your question. Then, it gets to work.

While it generates the report in the background, Deep Research autonomously browses hundreds of websites, sifting through the noise to identify relevant, high-quality information. You can even add more sources as it works, helping to guide its analysis.

But here’s the real kicker: The final report is “just the beginning.”

When the AI is done, it delivers a comprehensive, organized report that includes citations to all the relevant articles, websites, or papers it found. You can then, with a single click, import both the finished report and all of its sources directly into your notebook.

This is a massive time-saver. The AI doesn’t just give you a summary; it does the tedious legwork of building your library for you. That new, AI-curated collection of sources is now part of your NotebookLM knowledge base, and you can use all of NotebookLM’s other features—like quizzes, flashcards, and Audio Overviews—on the new material.

While Deep Research is the headline, Google is also making NotebookLM more practical by expanding the file types it can handle.

You can now link Google Sheets and ask for statistics based on the structured data. Instead of manually scanning a spreadsheet, you can just ask, “What was the average quarterly growth?” or “Summarize the key findings from this data.”

The update also adds support for:

• Google Drive URLs: You can now add files like PDFs via a URL link instead of having to upload them from your computer.
• Microsoft Word: NotebookLM will finally accept Microsoft Word files in .docx format, a long-requested feature.

This makes it much easier to dump all your project-related files into one place—spreadsheets, drafts, PDFs, and web links—and have the AI make sense of all of it together.

This update effectively bridges the gap between your private information and the world’s public information. NotebookLM is no longer just a smart notebook; it’s becoming a true research platform that can build its own knowledge base, saving you from the cognitive load of a dozen open browser tabs and helping you get to the actual “thinking” part of the work.