NotebookLM, which is Google's AI-powered research assistant, is one of the few AI tools that has earned a permanent spot in my daily routine. Though I do use the tool for my work-related tasks here and there, I'd be lying if I said my main use isn’t still studying. Given how often I use NotebookLM, and the fact that work makes up a significant chunk of my day, it only makes sense to find ways to fit it into my professional workflow too.

Lately, I've been pairing the AI tool with other productivity tools I use, like Perplexity, Excel, Apple Notes, and more. Then it hit me: I've been pairing NotebookLM with just about everything (even tools I don't necessarily open every day), but I somehow never thought to pair it with the one tool I open and rely on every single day: Slack.

I use the communication platform to connect with colleagues from all over the world, across all the teams I contribute to, and it's the one place where everything from quick updates to major project discussions happens. So, I did exactly that: I paired NotebookLM with Slack to find out if integrating my go-to AI tool with my most-used communication app could actually make a difference.

NotebookLM helps break down messy conversations into clear takeaways

From Threads to takeaways

NotebookLM has a lot of great features, but my favorite has always been its source-grounded nature. Like other AI chatbots, you can ask NotebookLM any questions you may have. But unlike most, it only pulls information from the documents you've added. This means its responses stay relevant, accurate, and rooted in your own material. I use this feature all the time to revisit Slack conversations and make sense of longer threads without having to read through every message myself.

Of course, given that it's an AI tool at the end of the day, I avoid using NotebookLM when it comes to sensitive conversations that involve personal details, confidential company information, or anything I wouldn't feel comfortable pasting into an external platform. For instance, if a few of my editors had a 40-message thread discussing an article someone pitched, I’d absolutely upload that. It’s helpful, low-risk, and saves me from scrolling through each message one by one.

Though I'm on five different Slack workspaces right now, I don't own or manage any of them. This means I can't directly export a Slack conversation as a .txt file. Even if I did have workspace owner or admin permissions, I'd need to export the entire workspace’s conversations and then dig through to find the relevant thread, since Slack doesn't let you export individual conversations.

Given that I typically use NotebookLM to analyze conversations that aren’t extremely long, I simply drag over the relevant Slack messages, copy them manually, and paste them into a fresh Notebook. It takes a few seconds, and once it’s in, NotebookLM can instantly start answering questions or generating quick summaries based on what’s there.

The issue with doing this using another AI tool like ChatGPT or even Gemini is that they often step in to share their own opinion or introduce information that wasn’t part of the original conversation. That can be useful in some cases, but when I’m just trying to understand what was actually said, not what the AI thinks was said, it becomes more noise than help.

NotebookLM helps me revisit what was actually said in Slack huddles

Making sense of what I missed

Though most of the calls I have throughout the week are on Google Meet, I attend a fair number on Slack Huddles too. While Gemini makes taking meeting notes on the former incredibly easy, things aren't as simple with the latter. Slack has an AI Huddle Notes feature too, but it can only be used if a workspace owner or admin has enabled Slack AI for the workspace. Unfortunately, this isn't the case for any of the workspaces I'm in. And since I'm not an admin in any of them either, I can't turn it on myself.

I haven’t been able to find a solid AI tool that works directly within Slack Huddles, so for now, I rely on workarounds. What I’ve been doing lately is recording the meeting to get a transcript. I then upload the transcript to NotebookLM, where I can ask it to make meeting notes, summarize the discussion, or simply pull out action items.

Again, since NotebookLM is source-grounded, it'll only refer to the transcript I provided to answer any queries I may have. It’s not a perfect setup, but given the limitations, it’s been a reliable and surprisingly efficient way to turn Slack Huddles into something I can revisit and act on.

Sometimes, I rely on NotebookLM to help distill long updates

NotebookLM comes in handy the other way around too

Given how fast-paced my work can be at times, there are moments when I want to quickly get the gist of something before passing it along. Maybe it's a lengthy press release that doesn't seem to end, or multiple articles I'm trying to find a pattern across. In those cases, I upload them to NotebookLM and ask it to break things down. Sometimes, I might simply ask it to highlight recurring themes, pull out the most important stats, summarize the key specs, or tell me the “wow” factor of the sources I provided. It gives me peace of mind that I’m not missing anything major before I pass it along to a team or editor.

Since I’m usually relaying the message further, I just can’t send something ahead without vetting it myself. AI has the habit of hallucinating after all, and you can never be too careful. Thankfully, NotebookLM includes citations right next to any claims it makes, and hovering over them shows you the exact part of the transcript it pulled the information from. This makes it so much easier to double-check what was said, especially when reviewing complex documents. It’s a huge time-saver, especially when I’m juggling multiple pieces of content and need to extract the essentials fast.

This pairing turned out better than I thought

I'll be honest — I had my doubts about pairing Slack and NotebookLM. Unlike with Perplexity, Apple Notes, and the other tools I've paired with NotebookLM, pairing Slack with it was a much different experience. But all in all, I'm glad I gave it a shot, because it's changed my workflow in small but meaningful ways.