In this episode, we’re diving head-first into the labyrinth of marketing with a trusty AI sidekick. If you’ve ever sat through a 60-minute meeting and wondered how to distill the chaos into a concise email summary, look no further.
Watch as we transmogrify meeting transcripts into clean, actionable summaries with the power of AI tools like Claude Pro, Sonnet 4.5, and more. You’ll learn how to navigate the sometimes murky waters of team-shared AI tools and come out looking like a marketing rockstar. Oh, and keep an eye out for those pop culture Easter eggs a la GenX sprinkled throughout!
[00:00:00] Introduction and Caution
[00:00:18] Overview of the Video’s Purpose
[00:00:43] Introducing Claude AI
[00:01:03] Setting Up Claude for the Task
[00:02:01] Uploading and Processing the Transcript
[00:02:50] Reviewing the AI-Generated Summary
[00:05:14] Customizing the Summary
[00:06:07] Final Tips and Best Practices
[00:09:05] Conclusion and Next Steps
[00:09:40] Outro and Call to Action
Introduction and Caution
[00:00:00] If you’re using an AI tool that is provided by your company, and if you’re in any way using something that might be shareable or accessible by other team members, don’t do stuff like this.
Overview of the Video’s Purpose
[00:00:18] This video is going to show you how to take a transcript from a meeting and work with AI to develop that into a really good email summary that you can distribute to that team. That includes next steps, quick summaries of everything that was discussed, and gives everybody the short and sweet of what you just talked about for 60 minutes. So now you can go on to the next thing and get your work done.
Introducing Claude AI
[00:00:43] I will admit to you right now that I did ask Claude, and Claude is what I’m going to be working with today. I use Claude Pro, Claude AI. I worked with Claude, and I said, “Claude, I’m going to have you fabricate a meeting transcript for me.” You’ll see as I open it that it had some fun with it. Okay?
Setting Up Claude for the Task
[00:01:03] I’m going to be using Claude for this exercise, and we already did a video about what LLM models or modes to use. I went over that in a prior video, so please do check that out. Because I want some extended thinking in here, I have that toggled on, and I’m going to be using Opus 4.6. Actually, I’m going to use Sonnet 4.5, best for everyday tasks and fastest for quick answers. Okay, so I am telling Claude, and again, Sonnet 4.5, what I need, and it’s generating its response here.
[00:01:36] I’ve never been on a submarine, though I’ve watched submarine movies. In submarine movies, they’ll call out a command and the command is repeated back. It’s confirming what you’re asking for. A lot of times your AI—whether you use Claude, Gemini, ChatGPT, Perplexity, or Copilot – will come back and say, “Great. I think I understand the task. Is this correct? Let me know if there’s anything else. Otherwise, let’s get started.” That’s essentially what it’s telling me here. “Drop the transcript when you’re ready.”
Uploading and Processing the Transcript
[00:02:01] So I’m going to go ahead and upload the transcript that I downloaded. Again, you can get a transcript if you’ve recorded the meeting. If you have not recorded the meeting, but maybe you have used the AI companion, you might only be subject to what the AI companion gives you, along with your own memory and notes.
[00:02:20] So… All right. Here’s the transcript, and I’m not even going to type in “Here’s the transcript.” I’m just going to upload the text file. Now it’s thinking about and reading through that entire text file of the transcript, and it’s going to process what I’ve got. Disclaimer: I had Claude help me develop the transcript because I’m in some meetings that I’m not going to be sharing with you since I do sign things like NDAs. So you’ll notice here in a moment that everything is quite fake.
Reviewing the AI-Generated Summary
[00:02:50] So we’ve got a Q2 Horizon Launch Campaign Planning Sync. The bottom line is that it’s telling us the timeline and key milestones, budget reality check, influencer strategy, and paid social. If you are a Gen Xer, you’ll understand a lot of these names. There’s Dana Scully from The X-Files, Jim Halpert – I believe he’s from The Office – Leslie Knope from Parks and Recreation, Peter Venkman from Ghostbusters, Ellen Ripley from Alien, and Marty McFly from Back to the Future. Notice that it took the entire text file.
[00:03:23] Real quick, by clicking on this, I can actually show you what the source file is. With this, you can see it’s a Zoom AI Companion Meeting Summary. It’s giving you the basics: how long the meeting was, what date it took place, the host, the participants, an overview, and the key topics discussed. This is very much like what an AI companion can give you, but then we’ve got action items and things like that.
[00:03:45] But then we’ve also got the whole transcript. I tend to like working with a transcript more than an AI companion because it’s working from whatever it could audibly process from the audio tracks of the recording. It’s able to say who we’re talking about, but it also means it may not be perfect. Remember, AI companions don’t necessarily go validate how someone’s name is spelled by who attended the meeting. It’s really just taking it from the audio. This gives you an idea of what the transcript was, including the time codes of what was happening throughout the discussion. Okay, so I’m going to close that for right now.
[00:04:25] Here we have the synopsis. There were timelines and key milestones, a budget reality check, and even talking about what was discussed. “Dana is exploring options for additional funding,” “Worst case, they’ll scale back the video,” “Although Ellen has strong feelings about that,” “action items by owner,” and it’s giving nice little things like TBDs and what have you. So that’s how Claude interpreted [the meeting]. It said what decisions were made, whether there were blockers or dependencies, what things still needed clarification or put in the parking lot, and the next meeting. Then there’s even a pro tip for Marty: “When you send that recap email, you might want to flag the budget gap in red.” So it’s giving me guidance, saying, “Hey, you might wanna, as you’re preparing this for everybody to send it around, think about this.”
Customizing the Summary
[00:05:14] Now, did this give me exactly what I’m going to send around to the team? Probably not, because there are still some TBDs here. You also refine this. I can say: “This is great, but I actually would like it to be more succinct. I’d like the next steps or action items to be up top, the first thing people see, and I’d like to make sure we’re including enough deference and respect for Dana Scully since she’s the team lead. She tends to care about things like that because she’s likely to forward this to her VP.”
[00:05:56] All right. I completely made that up, but I wanted to show you that you can massage this a bit. This is where prompts come in. You can give it whatever context you want.
Final Tips and Best Practices
[00:06:07] I want to add a caveat here. If you’re using an AI tool that is provided by your company, and if you’re in any way using something that might be shareable or accessible by other team members, don’t do stuff like this. It’s better to maybe take it off and do your own account with an AI tool. I say that because a lot of companies, wanting to make sure they’re embracing AI, are probably setting up team accounts. I’ve been on a client account where there was one login that multiple people could use to get into to use ChatGPT. Well, that meant we could all see each other’s chats. So you certainly wouldn’t want to go in there and say, “Help me work on my résumé.” No, and don’t do that anyway. You’ve got to keep work stuff and personal stuff separate when you’re talking about what systems you’re using. But keep that in mind. Use your prompts to get what you need out of these AI tools. But maybe you need to not censor yourself a whole lot; you might want to use something on the side that is not connected to your work and cannot be accessible by anyone at work, so that I can be honest about what I need to output. That might be good guidance for you.
[00:07:20] But notice that it’s got action items first. Dana Scully, the team lead, for the “ego” example I mentioned earlier, completely made up, but she’s got her stuff up front and more prioritized than everybody else’s. Then we have the campaign overview, budget status, all the decisions and dependencies, and everything being approved by Dana. So again, we’re giving her some deference, and it puts Dana’s leadership front and center. So that’s an example.
[00:07:45] Again, you can take a text file, a .doc or .docx file, or any kind of file. You could even take a PDF. Or, if you’ve got the transcript, you can simply copy and paste it directly into your chat; it can work with that. You can upload those files [into the chat], and then [the AI] can give you the output you need and massage that a little bit. It can save you so much time because previously, without the help of AI, you would be, “Oh, I better take notes this entire meeting because I know I’m going to have to turn this into a summary later,” “Oh then, I might have missed something because I also had to interact as a team member as well. I’m not just a receptionist or a note taker; I actually contribute to this team too. I missed something.” Working with AI is much more like working with someone who was in the meeting, who was taking notes, who has the transcript, and who can also say: “Oh, and by the way, don’t miss this.”
[00:08:41] One final note: Do make sure, don’t just copy and paste the AI’s output without reviewing it. Make sure it says and captures every single thing, and you can actually give that feedback to the AI and say: “I appreciate this, but I made an adjustment. You missed this part. Let’s make sure we don’t miss that going forward.” It will actually take that feedback into account.
Conclusion and Next Steps
[00:09:05] All right. So that was how to use AI for one of the more tedious tasks that many of us run into. Especially when we work on lean teams, this helps keep everybody moving along and staying productive. AI can absolutely help you get that done. So use this, try this out the next time you’re in a meeting. In a future video, I’m going to show you how you can do—yes, we can produce this document, but we can also create a deck out of this if that’s what you would like and what would make Dana look best to her superiors. Coming soon in the next video. Thanks for watching, and let me know what you think.
[00:09:40] You’re still here? It’s over! But before you head out, hit Like on this video, and subscribe if you want to get “Moore” videos that’ll help you realize how much “Moore” there is to marketing. See what I did there?
Exclusive content is available with Rachel’s livestream videos: Boss Level Mondays and Hot Mic Fridays. Start your week off with irreverence and real talk as you face down the work week of a marketer with Rachel at your side. Then, on Fridays, head to Hot Mic Fridays for unfiltered takes from the week, as well as podcasting and video tips you can always use. Subscribe here on Substack to access and join Rachel live!