NotebookLM has become one of my favorite tools for turning documents, research papers, and notes into something I can actually understand and use. But as useful as it is, I have been looking for an alternative that fits more naturally into my everyday workflow. Claude Projects came close, but it still felt more like an organized chatbot than a complete research workspace.

Then I started using Copilot Notebooks, and it felt like the productivity hub I had been trying to build for myself.

Moving past the two-player hype

The sleeper hit

For a while, the conversation around AI-powered research workspaces felt like a two-player contest. NotebookLM was the obvious choice for working with sources, while Claude Projects was the alternative people recommended for more flexibility and a stronger general-purpose assistant.

I bought into that comparison, too. Whenever I thought about organizing research, uploading documents, or keeping an AI focused on a specific project, those two were the tools I considered.

Then I came across MS365 Copilot Notebooks on a Reddit thread, and that changed the comparison for me. What I found was that Microsoft had been quietly playing an entirely different game.

While everyone else was arguing over Google versus Anthropic, I have quietly moved my entire research and writing workflow over to Microsoft’s sleeper hit.

The Microsoft ecosystem is its biggest advantage

Fits naturally into the way I already work

The biggest reason Copilot Notebooks works so well for me is that it doesn’t feel like another AI tool I need to force into my routine. I have already spent a large part of my day inside Microsoft’s ecosystem, so the notebook has become an extension of the tools.

OneNote is a good example. I use it almost every day to capture article ideas, save research notes, organize rough outlines, and keep track of information I may need later.

Since I can create and access my Copilot notebooks from OneNote, I don’t have to leave that familiar environment whenever I want AI assistance. My regular notes and AI-powered project spaces sit much closer together, which makes the entire workflow feel more integrated.

The same applies to Word and Excel. I use Word for drafting and editing articles, while Excel remains useful for organizing and tracking information and working with structured data.

When I start a Copilot notebook, I can pull in those files directly from my OneDrive account, along with any relevant PDFs. I don’t have to download everything locally, upload it to another service, and then repeat the process whenever a file changes.

The notebook becomes the layer that connects all those materials and helps me make sense of them.

I can also create new pages directly inside a Copilot notebook when I need to write something that does not belong in an existing file. The editor feels familiar because it offers the same robust writing experience I was used to in Microsoft Loop. I spend less time moving information between services and more time actually working with it.

It’s way better than Notebook’s standard notes, and there is no such option to create notes in Claude Projects. I have to rely on a markdown editor to get the job done.

Copilot Notebooks is packed with features

All the essentials are covered

Copilot Notebooks may not have a long list of features, but it handles the essentials like a pro. Once I add my sources, I can ask questions about the project and receive answers grounded in the files I have provided.

The responses can be slow, especially when the notebook contains several large documents, but I have generally found them to be accurate enough to justify the wait.

I also like the way Copilot structures its responses. It uses headings, bullet points, and emojis to separate different types of information. The emojis may appear like a small cosmetic touch, but they make longer answers easier to scan and understand.

Instead of receiving a dense wall of text, I can quickly identify costs, recommendations, warnings, and next steps.

I experienced this while creating a notebook for a home renovation project. I added the relevant PDF estimates, Word documents, and Excel sheets detailing materials, contractors, and project expenses.

When I asked Copilot about the painting costs, it didn’t simply pull out a single number. It also surfaced related information, such as the paint company, the type of paint mentioned, and other details related to that expense.

I could then take the conversation further by asking how to reduce the renovation budget. Copilot examined the available information and suggested detailed, practical ways to lower costs, such as reconsidering certain material costs or adjusting parts of the plan.

Because Copilot had access to the notebook’s sources, its suggestions were more relevant to the actual expenses I was dealing with.

However, the experience wasn’t without limitations. Sometimes, Copilot Notebooks take ages to load on the web, and responses can be painfully slow for normal questions.

Second, there is no way to add an entire OneNote notebook as a source. I need to copy each page link and manually add it as a source, which is time-consuming.

I would also like to see more features in future updates to bring it on par with Google’s NotebookLM.

The ultimate AI knowledge base tool

What makes Copilot Notebooks more useful to me is not a single standout feature, but how naturally it fits into the Microsoft ecosystem I already use every day.

Instead of moving information between separate tools or rebuilding context in another chatbot, I can stay closer to the apps and services where my work already lives.

Look, NotebookLM remains an excellent research tool, and Claude Projects is still powerful, but Copilot Notebooks feels more practical for my routine.