NotebookLM Is Now Gemini Notebook: What Changed and What to Do

Google renamed NotebookLM this week. As of July 16, it is Gemini Notebook, per the Workspace Updates announcement, and before anyone braces for a migration: your notebooks stay, your shared links keep working through automatic redirects, and your IT department has nothing to do. This is a rename, one of the least disruptive kinds of change Google ships, and also one of the most confusing to talk about for the next month.

Here is what changed, what only looks changed, and the short checklist worth running if your team leans on shared notebooks.

A rename, not a rebuild

The announcement is unusually clear about scope. NotebookLM becomes Gemini Notebook. Automatic redirects preserve shared notebooks, so the link a colleague bookmarked in March still resolves. No admin action is required. And the rollout may take more than 15 days, which means the name on your screen and the name on your coworker’s screen may disagree into early August. That is expected, not a bug.

Everything that made the product what it is survives untouched: the source-grounded design, your uploaded documents, the notebooks themselves, the material you have generated from them. If you walked away from your notebooks on Tuesday and came back to a different name on Thursday, the correct amount of concern is none.

The one genuine to-do the rename creates is verbal, not technical. Any internal documentation that says “put the files in NotebookLM” now points at a name that is disappearing. Redirects fix links. They do not fix the new hire reading a runbook that references a product they cannot find in a search of Google’s homepage. A five-minute find-and-replace across your team’s onboarding docs closes that gap.

The change that actually mattered came in June

The rename is the visible event, but the functional story happened a few weeks earlier. In June 2026, Google upgraded NotebookLM, as it was still called, with advanced reasoning and code execution, per the company’s AI updates roundup published July 1.

Those two capabilities move the product up a weight class. The original pitch was always source-grounded answers: upload your documents, ask questions, get responses tied to what you uploaded rather than the open internet. Advanced reasoning means the answers can do more work, following a chain of implications across sources instead of retrieving the nearest relevant passage. Ask how the assumptions in a January planning doc hold up against the numbers in a June report, and the model has a better shot at actually reasoning through the collision rather than summarizing each document at you.

Code execution is the quieter one, and for office workers it is arguably bigger. It means the system can run computations rather than estimate them. Totals, comparisons across a spreadsheet you uploaded, anything where the honest answer requires arithmetic instead of pattern-matching. Language models are famously shaky at math done in their heads; letting the model write and run code to get the number is how that gets fixed. You do not see the code. You see answers about your data that hold up when you check them, which was not consistently true before.

A concrete way to feel the difference: upload a budget spreadsheet and the memo that justified it, then ask whether the spending in the sheet actually matches the plan in the memo, category by category, with the discrepancies totaled. That question needs both capabilities at once. The reasoning half connects the memo’s intentions to the spreadsheet’s line items; the execution half does the arithmetic on the gaps. Before June, you would have gotten a plausible-sounding narrative answer. Now you can get one with numbers that survive a spot check, though you should still run that spot check, because grounded and computed does not mean infallible.

If you tried the product a year ago and filed it under “interesting for summarizing PDFs,” the June upgrade is the reason to re-file it. The rename just makes this an awkward moment to rediscover it, since half the coverage you will find uses the old name.

If your team shares notebooks, run this checklist

None of this should break anything. “Should” is doing some work in that sentence, so a short verification pass is cheap insurance for teams that treat shared notebooks as working infrastructure:

1. Open your most-used shared notebook from an old bookmarked link and confirm the redirect lands where it should.
2. Check that everyone who had access still has access, especially collaborators outside your immediate team.
3. Search your wiki, onboarding docs, and pinned messages for “NotebookLM” and update the name while the old one still rings a bell.
4. If a notebook link is embedded somewhere brittle, a client portal, a slide deck template, a printed QR code somewhere in the office, test that path specifically rather than assuming.

Fifteen minutes, most of it step three. And since the rollout may exceed 15 days, do not treat a teammate still seeing the old name as a problem to escalate. The name arrives when it arrives.

One more habit worth adopting during the transition window: when you share a notebook link in chat or email over the next month, label it. “Here’s the Gemini Notebook (formerly NotebookLM) for the Q3 project” costs you six words and saves the recipient a confused minute wondering whether this is the tool they already know or something new they are expected to learn. Renames create exactly one kind of real damage in organizations, which is people quietly not clicking things they do not recognize, and a parenthetical fixes it.

Why Google is doing this

Google has spent two years consolidating its AI products under the Gemini name, and NotebookLM was one of the last major holdouts still carrying its research-project branding. The “LM” in the old name stood for language model, which described the technology in 2023 and describes almost every product Google ships in 2026, making it useless as a differentiator.

Gemini Notebook tells you what the thing is and whose family it belongs to. For Google, that is tidier marketing. For users, the practical upside is discoverability: when the product sits inside a naming scheme people already know, finding it, asking about it, and getting your company to approve it all get slightly easier. The cost is a month or two of “wait, is that the same thing?” in meetings, plus a long tail of tutorials, YouTube walkthroughs, and internal docs that reference a name Google no longer uses. Search for both names when you are looking for help; the old one will dominate written guidance for a while.

There is a mild lesson here about betting on AI tool names in general. Products in this space get renamed, folded into suites, and reshuffled between tiers with some regularity, and the teams that suffer least are the ones whose processes reference what a tool does rather than what it is called this quarter.

What to actually do

For most people: nothing. Keep using your notebooks, expect the new name within a few weeks, and correct colleagues gently. For teams with shared notebooks embedded in real workflows, run the four-step check above once the rename reaches your account. And if the June reasoning and code execution upgrades passed you by, spend an afternoon pointing the newly renamed Gemini Notebook at a real pile of project documents and asking it questions that require actual computation. The name is the least interesting thing about what this product became this summer.