Gemini moves toward project work
Google rolled updates that fold NotebookLM into Gemini as persistent 'notebooks', add scheduled actions for recurring prompts, and let Gemini produce interactive 3D models and simulations—moves that shift the tool from one-off chat to ongoing project support. Those features tether AI outputs more tightly to project context and cadence, which matters for teams needing repeatable deliverables rather than ad‑hoc answers. (techrepublic.com) (androidauthority.com) (tech.yahoo.com)
Google is turning Gemini from a chat box you visit once into a workspace that keeps your files, remembers the project, and shows up again on a schedule. On April 8, Google said Gemini now has “Notebooks,” a feature tied directly to NotebookLM, the company’s document-grounded research tool. (blog.google) A notebook is basically a project folder with a brain attached. Google says you can put research, notes, and chats in one place, then move between Gemini and NotebookLM without rebuilding the context each time. (blog.google) NotebookLM already worked like an assistant that reads only the sources you hand it, instead of guessing from the whole internet. Google’s Workspace site describes it as an “AI research and thinking partner” built around the information you trust. (workspace.google.com) That changes the rhythm of using Gemini. Instead of asking one question, copying the answer, and starting over tomorrow, a team can keep the same notebook alive as a running brief for a product launch, a class unit, or a sales account. (blog.google) Google has also been adding scheduled actions, which means Gemini can run the same prompt again later without being asked from scratch. Google’s help page says users can set up recurring actions in the Gemini app, edit them later, pause them, and keep up to 10 active scheduled actions at once. (support.google.com) Google first announced scheduled actions for the Gemini app on June 6, 2025, pitching them as a way to handle routine tasks and deliver personalized updates automatically. That makes Gemini look a little less like search and a little more like a junior staffer with a calendar reminder. (blog.google) The third shift is visual. Google’s Gemini updates page says the app can now generate interactive simulations and models, so the answer is no longer just a paragraph or a picture but something you can rotate, inspect, and adjust. (blog.google) That matters because some work is easier to understand when the answer behaves like an object instead of a sentence. A molecule, an orbit, or a mechanical system makes more sense when you can drag it around and change the inputs than when you are staring at static text. (blog.google) Put those three pieces together and the pattern is pretty clear. Notebooks give Gemini memory tied to a project, scheduled actions give it cadence tied to a calendar, and interactive models give it outputs that people can explore instead of just read. (blog.google) (support.google.com) (blog.google) Google is not just trying to make Gemini answer better questions. It is trying to make Gemini stay with the work after the first answer, which is where most office software actually earns its keep. (blog.google)