Excel can be really intimidating to use. There, I said it. I’m not talking about creating a basic spreadsheet or formatting cells. I mean using Excel as a proper data analysis tool. The kind where you're expected to use complex formulas, manually figure out trends, and build something that actually makes sense. It's a lot, and it's a task I've never been a fan of doing.

At the same time, it’s something I can’t avoid. Whether it's for a college project or work, Excel always finds a way to sneak back into my workflow. So, instead of avoiding it altogether, I decided to bring in some help: NotebookLM.

To date, NotebookLM is my favorite AI tool, and it's actually managed to convince me that AI can be a game-changer for productivity. I recently realized that while using it alone already has its strengths, pairing it with tools like Excel can unlock an entirely new kind of workflow.

I upload my Excel spreadsheets to NotebookLM once the data starts to blur

Turning spreadsheet chaos into clear insights

NotebookLM isn’t a spreadsheet tool, so I can’t really use it to “replace” Excel. I still need to manually do the work to create my spreadsheets, clean up data, and apply formulas. So, I begin the process inside Excel like I normally do. I create a new spreadsheet, add in my data, and organize it into tables just to make it easier to read.

Where my process has now changed is in what I do after the spreadsheet is ready. Instead of spending hours trying to make sense of it all on my own, I upload the file to NotebookLM and let it help me break things down, find patterns, and make sense of the data faster.

Unfortunately, NotebookLM doesn't currently accept .xlsx files as a source. So, I export my spreadsheets as PDFs and upload them to either an existing or new NotebookLM notebook. If you want to, you can also upload a .xlsx file as a .csv, .txt, or Markdown document.

NotebookLM helps summarize insights and surface trends

Making sense of your data faster

As I mentioned above, I'm using an example spreadsheet for this one due to confidential information in the sheets I typically use. The example spreadsheet had 5,000 rows, which is a good representation of the kind of large datasets I usually work with, and also a lot to go through manually. It's exactly the kind of spreadsheet I'd rather let NotebookLM analyze for me, simply because of how much manual labor it'd take otherwise.

As soon as you upload a source to a NotebookLM notebook, it summarizes it in a paragraph right below the name of the notebook (which is automatically generated based on the source content). It also generates three questions, which I find are a good starting point to understand the data better, especially when I'm not sure what to focus on first. For instance, here are the three questions NotebookLM suggested for the 5,000-row spreadsheet:

• What are the demographic distributions across gender, age, and country?
• How do numerical trends, such as IDs, vary throughout the dataset?
• What patterns exist in the recorded dates for individuals within these sources?

I decided to ask NotebookLM all three questions, and the answers were all incredibly helpful. For instance, in the first answer, it pointed out that while the spreadsheet had 5,000 rows, the data actually consisted of 50 unique entries repeated 100 times. It then broke down the demographic distributions across gender, age, and country, calculated the average age, and even grouped individuals into age brackets.

Similarly, in the second answer, it analyzed how the IDs were structured across the dataset. It pointed out that most of the spreadsheet followed a sequential, ascending ID pattern. Interestingly, later in the file, NotebookLM noticed a shift: a set of non-sequential, repeated IDs that didn’t follow the earlier pattern.

Finally, in the third answer, NotebookLM flagged that all entries were tied to just three specific dates that repeated throughout the sheet: 15/10/2017, 16/08/2016, and 21/05/2015. Every time you ask a query, three more follow-up questions appear, which can be helpful when you're not sure what to dig into next. Of course, you can ask any questions you may have about the spreadsheet you uploaded, and NotebookLM will pull the answers from your data within seconds.

NotebookLM can create a detailed report of your uploaded dataset

Reports that save you hours

It’s not uncommon to be asked to make a report of a massive spreadsheet. And no matter how many times you’ve done it before, it can still feel overwhelming, especially when you’re staring at thousands of rows that just don’t seem to end. NotebookLM makes that process a lot easier. And though you don’t necessarily have to copy-paste the exact report as-is (I wouldn’t recommend that either), it’s a great starting point that saves a ton of time.

For instance, I prompted NotebookLM to generate an in-depth report about the dataset, and the report it generated included:

• Dataset Overview and Structure
• Demographic Distributions
• Numerical Trends in IDs
• Patterns in Recorded Dates
• Overall Insights

The report included relevant examples and had citations throughout, which made it really easy to verify everything quickly. Even if I wanted to tweak the wording or structure later, having that initial outline meant I wasn’t starting from scratch. It gave me a clear sense of what insights were worth highlighting and saved me the effort of digging through the sheet manually to spot trends or anomalies.

NotebookLM rarely hallucinates, so you can actually trust the answers

Trusting NotebookLM when accuracy matters

This isn't my first time resorting to AI tools to help me out with spreadsheets. However, whenever I do, I’ve noticed that there are times when the answers feel a little off, like the AI is guessing or pulling info that doesn’t actually exist in the file. That’s always been my biggest concern when using AI for anything data-related. Now, I'm not going to claim that NotebookLM’s answers will always be right. The tool itself mentions, "NotebookLM can be inaccurate; please double-check its responses."

However, I’ve been using the tool since before it even rolled out widely, and I can't recall a single time it gave me an incorrect answer. The best part about NotebookLM is how easy it makes it to verify the information it gives you. Citations are always displayed right next to each response, so if you’re unsure about something, you can just click and jump straight to the source. That’s been super helpful for cross-checking details without having to manually scan through thousands of rows.

Excel and NotebookLM are a solid team

If you're like me and the mere thought of Excel and spreadsheets gives you nightmares, you need to try pairing it with NotebookLM. Its free version is powerful enough to tackle large datasets, so you don’t even need to pay anything!