Millions of people have turned to NotebookLM since it launched three years ago for studying, synthesizing and summarizing documents, and data organization. Now, NotebookLM is getting a big upgrade to become an even better research assistant. 

Google said its AI product will gain "new agentic capabilities in chat and more advanced reasoning" to handle even more complex research tasks and projects. Below, I'll break down the new updates coming to NotebookLM, rolling out globally now to Google AI Ultra subscribers and Workspace business customers with AI Ultra access. 

Read also: Gemini Gets New Notebooks Feature That Syncs With NotebookLM

Chat gets a major upgrade

NotebookLM's chat tool now runs on Gemini 3.5 and Antigravity to improve accuracy and give clearer visibility into its reasoning and outputs, which Google said was a highly requested feature. 

Each notebook includes a secure cloud computer, so NotebookLM can write and run code for deeper research and analysis. NotebookLM also has more than 100 curated software skills to expand its thinking steps and allow it to create a wider range of results.

More formatting options

NotebookLM can now output answers in a wider variety of ways. The AI tool will assemble context from your sources into downloadable, editable artifacts, including data visualizations (png, svg), documents (PDF, DOCX, Markdown, text), images via Nano Banana (png, jpg, gif), structured data (CSV, JSON) and Microsoft formats (XLSX, PPTX). 

Google plans for more output formats to come in the future.

Research gets easier and more trustworthy

Getting started with research is easier. With the upgraded NotebookLM, you no longer need a fully formed repository of sources. NotebookLM can help you build one from loose ideas, guide source discovery inside chat and use Google Search to surface high‑quality web sources. You can also choose which sources to add to your research, and each source stays clearly attributed, so results remain grounded in information you trust.

Read also: NotebookLM's Video Overviews Just Got Better Thanks to a Trifecta of Google's AI Models

These changes to NotebookLM aim to broaden real‑world use cases. For instance, a researcher can merge messy international datasets, run code to analyze them and produce charts plus a PDF report; technical teams can convert dense specs into simplified guides and slide decks; and small business owners can combine sales and spending data, then use NotebookLM's analysis and reporting to inform expansion decisions.

All of these changes combine to help you make the most of your research experience, for better results and a more transparent process.