Transcription
Replexity isn't just some smart search engine anymore. With Comet, it can scan your inbox, summarize attachments, analyze your content, even write LinkedIn reports for you. You can create fully autonomous AI agents that execute entire workflows for you.
In this video, I'm going to be breaking down seven mind-blowing use cases of Perplexity Comet agents so that you can save hours of time. By the way, if you're new here, my name is Rick Mully and for the past 11 and a half years, I've run a multi-figure online business, helping other online business owners grow and scale their business. Today, I run a community called the AI playbook where I am an AI strategist and I help online businesses leverage AI so that you can become more efficient, streamline your business, and increase profit in the process. I'll link to the community in the description below.
All right, so here we are inside the Comet browser. You're going to see this right here. You also have the assistant in the upper right hand corner, which we are also going to be using in a couple of these use cases I'm going to be sharing with you. What I love about what you can do with Comet agents is that you can leverage what I like to call agent chaining. So, chaining agents together to accomplish a multi-step workflow. And that's exactly what we're going to do in this first use case. I know there's an email in my inbox. For this use case, I want it to go find the email. Number one. Step number two is find the link within that email. Step number three, I want it to click that link in the email. And then step number four, summarize it. And then step number five, give me that summary. Bring the summary right back here. So, I'm going to give it a super simple prompt. There's an email in my inbox from Michael Hyde. It includes a link to his customer avatar. Please give me a detailed summary of that customer avatar. So, I'm going to click go. It'll share the steps that it's taking to accomplish this task. All right. And really quickly, it's given me a detailed summary of the customer avatar. So, it read the link in the email, uh, consumed it, analyzed it. Now, it's given me the type of summary that I asked it for, and it was really, really fast. And so, again, it's chaining together multiple agents for this multi-step workflow. In this case, it was about five steps, and it did it really, really well.
All right, before we do use case number two, I want to show you a couple of quick things here. So when you click in the box here, it's going to give you some ideas. So like summarize all my events for this week into a quick agenda that I can scan. So again, you can see more examples of chaining agents together to accomplish an entire workflow. Now down the bottom here, they have these things called shortcuts. So these are like GPTs that you can bring up right here within Comet that you have a keyboard shortcut for, or you can just click on it and it will come right up. So you give it a name, you give it the instructions or what the the prompt is that you want it to do. You can choose a mode down here. You can choose the model that you want to use for it, which I like a lot. You can choose the sources just like you could in a normal perplexity search. And then click save.
For this next use case, we're going to transform our content strategy. I want it to scan my YouTube channel. I've given it the link to my YouTube channel. I'm telling it to break down video performance by topic and recommend new trending topics based on niche analysis. And again, you can see the agent working through. You can see here for me it's found my channel and this is going to take a little bit longer because it's going through 64 videos and analyzing all of them. But I have no problem with it taking a while. Number one, it would take me a ton of time to do this manually, right? All right. And after a few minutes, it's given me a comprehensive breakdown of my channel uh by performance, by topic, top performing videos and topics, underperforming topics, and then it gives me trending niche topics and recommendations for what I could be talking about next. and it's super helpful and actionable.
Okay, for use case number three, we're going to build a news concierge agent. This use case comes directly from one of our members of the AI playbook community. He has a weekly newsletter as it pertains to money and finance and he does a ton of manual Google research to find the types of stories that he wants to talk about in his newsletter. So, why not create a Comet agent to do all the news gathering for you? Here's the prompt that I've given it. Act as a world-class helpful research assistant. I want you to help me in find select articles. I'm giving it criteria. I'm even linking to so here in comment I can just use the at symbol and I can link to other tabs or reference other tabs as context. So I've linked to a specific article that my community member is looking for and then I've told it what kind of sources to look for and then also the kind of output that I wanted to give me. So let's click go. All right. After several minutes I finally have the results and as I scroll down here you can see uh let's see. Should you replace your financial adviser with AI? AI is transforming how we bank, what this means for your money. These are perfect. These are exactly the types of things that I'm looking for. It just saved a bunch of time doing manual research in Googling and so forth. So, now that I know these instructions work really well, I can save the query as a shortcut. I can click on this button here and I can call the shortcut weekly AI money news. I can come down here, click on whatever model I want. I can click on different sources. And so all I have to do is just literally click that shortcut next time and it will do this exact search. Now the other thing that I can do with this is I can create a task of this agent. So I can click on create task and then I can schedule this. So if I want to be doing this weekly for example like every Friday morning, weekly Friday morning at 8:00 a.m. notification platform. So I can have it texted to me. I can have it emailed to me. I can have it sent to me within the Comet app here.
All right, for use case number four, let's create a LinkedIn content researcher agent. So, the reason that this is important to me is I wanted to analyze a list of given LinkedIn accounts and then compile a report giving me the most engaged post for the last seven days across these, in this case here, five LinkedIn accounts. So, here's my prompt. You're going to be given a list of LinkedIn accounts. Please write a consolidated report on the 10 posts amongst these accounts that have received the most engagement, reactions, and comments in the last seven days. I give it some instructions on what to include. Then I want it to open my Slack and then add this report to a specific channel. Let's click go and see what this does for us. All right, it started to do its research and analysis. It's going through the post here. Looking in the preview window here, you can see that all right, it's on the it's in LinkedIn. It's on the accounts. All right. Now you can see here in the small window that is in my Slack now. All right, here we are in my Slack for Rick to review channel and here is the LinkedIn report. So again, that would have taken a ton of time for me to do myself. Comment just did that in probably 15 20 minutes. Pretty good.
For use case number five, we're going to build a 360°ree social intelligence agent and we're going to leverage the pre-done shortcuts. Here are all the shortcuts that I can choose from. Again, you can create your own, but I want to use a pre-built one that came with Comet, if you will, out of the box that is called Trending on Social. And then I'm going to add some additional instructions to this where I'm going to say, I want to know about the latest trending topic. Scroll through each of my social media websites that I'm logged into. What are the top discussed AI applications in my industry this week? All right, in probably 2 minutes, it's giving me a really nice report here. Here's a summary of latest trending topics, and it breaks it down by platform. So, LinkedIn, Twitter, Instagram, and it also gives a a really nice table down here. Now, while it was going, I opened up Tik Tok. I opened up a a tab here in the browser for Tik Tok. And so, let's just say that we want to add Tik Tok to this report. I'm going to say, let's add Tik Tok to this report. I've just opened. Okay. And it's added the Tik Tok top trends in AI, content creation, digital business. gone through topic by topic, giving me the source link directly to that specific Tik Tok video, which I think is really cool. And this comment agent just did it in a matter of minutes.
For use case number six, we're going to create a partnerships manager agent. So, I get a ton of collaboration requests for the YouTube channel here. And so I want to create an agent here in Comet that will go through my email inbox, find those collaboration emails, do a slight analysis on the request, and then put all the requests into a notion database that I can then go through really quickly and figure out if I want to take next steps on it. So here's the prompt. Find all emails about partnership opportunities in my Gmail inbox for the past 5 days. Add them to my notion database. I want it to extract the following information from the partnership emails that it finds. So, sender name, sender company. And by the way, these are the names of the columns I have in my notion database. And then in the notes section of the database, I'm telling it since it knows a lot about me, I want it to analyze it and tell me whether it thinks that I should pursue it. Then finally, I'm saying map the extracted information to the matching fields in the notion database. All right. So, as you can see here, it's going through my email inbox looking for those specific types of emails. And we're going to go over to Notion and try to watch it do this. And you can see here it just added nim.ide to a new row here. So it's it's working. Okay. It's just giving me an analysis on whether it thinks that I should pursue this collab. Highly recommend pursuing and it tells me why it recommends doing that. So I've just stopped the agent because there's a ton of collab requests in my in my email inbox. I'd probably want to schedule this agent to run on maybe a daily basis. So I can come up here to the three little dots, click on create task. Here are my instructions that I've given it. Then I can create the schedule. So let's just say I want to run this daily at um I don't know 7 or 10 a.m. And so now again, all I have to do now is go into my notion database every morning and I'll have all these collab requests in there and I can really quickly process them.
All right, for use case number seven, we're going to really take advantage of Comet's agentic capabilities. As we talked about early on, they have the assistant option here. So, when I click on this, a window pops up here. And we're going to use Comet's AI assistant here with OpenAI's new agent builder. Time recording this. This just came out yesterday. And what I think it's really, really lacking is the ability to use natural language to build out workflows in this agentic builder. So, we're going to use Comet's AI assistant to do just that. So, I'm going to click on create. So, I'm going to create a new workflow. This is kind of what it looks like if you haven't seen this yet. So, let's just say that I don't really know how to use OpenAI's new agent builder. And so, I'm going to use the assistant. I'm going to type in here. I want to be able to ask questions and get answers from the document in my file search. File search is over here on the left hand side. It's the file attachment called SEO is dead. Please build me an agent workflow here in agent builder to accomplish this. Notice here that it's just written the instructions for the agent all by itself. Right here, it named the agent. So again, it is doing this all by itself. Look at like my hands right here. It's I'm not doing any of this. But what I want to do now is I want to preview it. I'm going to ask um what is Ago? And you can see here it says I need to use the file search to answer the user's question. What is AGO? etc. Defining it. And it's going to give me the answer hopefully. Yeah, GEO optimizing. So your content AIO in short. And then also it gives you a link to the source file which is the SEO is dead text file. Now, how cool is that?
All right, for this bonus use case, I'm on a lot of podcasts as a guest to talk about AI and online business and so forth. So, I have this form and I want it to fill out this form for me based on the information that it knows about me. So, I'm going to type in based on what you know about me and my business, fill out this questionnaire to the best of your abilities. Skip whatever you are not sure about. And then you can see here it says, "I'll help you fill out this this podcast guest intake form." Yeah, you can see that it just filled in my name and it just filled in my bio all by itself. Filled in soapbox topics. Now, I will say that this is actually pulling old information. So, I would need to update this within the Comet browser. It's filling in my Instagram URL, my LinkedIn URL all by itself. Takes me no time. Obviously, I want to make sure that it has the proper and accurate information, but I love the fact that the AI assistant here in Comet can do this.
All right, so there you have it. Seven or eight with the bonus use case there. What I think are mind-blowing use cases of Comet agents and the type of things that it can do autonomously and thus saving you a ton of time and money. I tested Comet browser several months ago. I was not impressed with it. I'm very impressed with it and I'm now using it exactly like I've shared with you here today. Again, if you are an online business owner and you're looking to leverage AI in your business and you want AI strategies so that you can streamline your business, become more efficient, and increase profitability, I want to invite you to join me inside my AI playbook community, which I'll link to in the description below. So, thanks so much for watching as always and I'll see you in the next.