# 155 Perplexity Computer Masterclass for Beginners (full tutorial)
Ecom Podcast

# 155 Perplexity Computer Masterclass for Beginners (full tutorial)

Summary

The Corey Ganim Show shares actionable Amazon selling tactics and market insights.

Full Content

# 155 Perplexity Computer Masterclass for Beginners (full tutorial) Speaker 2: Perplexity Computer is insane. And in this video, we're going to bring you a true beginner's masterclass on how to use Perplexity Computer to replace hours of manual work with a few simple prompts. No code required, no expensive tools. So sit back, relax, and learn how to use this incredibly powerful tool. Eliot, what are we building today? Speaker 1: We are going to be building some SEO and local SEO machines using Perplexity Computer to just automate everything to do with your website SEO. Speaker 2: Awesome. What gives you the credibility to do that? What's your background? What have you done up to this point? Speaker 1: For just over 10 years now, I've been doing SEO, content marketing, and then ranking stuff at the top of Google. Deep in the AI world now, I do a lot of AI coaching as well on top of mixed in with the SEO stuff now. So basically, my whole day is working out how to do SEO faster, better with AI, which is what we're going to do today. Speaker 2: Awesome. Well, kind of give us a rundown, like what is Perplexity Computer? Before you jump into your screen, maybe explain that for people who haven't heard of it or maybe don't know how it differs from other tools. Speaker 1: Perplexity Computer is almost another evolution from things like Claude Cowork in that you tell it what you want, you tell it your outcome, and it goes and starts delivering. But it almost goes a step further than a lot of the other tools in that it runs in the cloud, so you don't have to run it locally on your desktop. Plus it can run, it has 19 LLM agents under the hood. So it can run multiple AI agents simultaneously in parallel of different LLM models. So we're basically getting the power of everything in one cloud container, which I'll show you why it's so powerful. Speaker 2: Yeah, let's see it. So let's go ahead and jump into your screen. Speaker 1: Let me jump on the screen share then. Unknown Speaker: Right. Speaker 1: Okay, so Perplexity Computer is kind of, this is the normal homepage to Perplexity, right? Looks like any other LLM model, but in here now we have this computer button and this changes the game in here because it gives you access to all those sorts of things I just talked about and it runs virtual computers on the back end. Now, to give you an idea of Almost yesterday when I started testing this stuff more in depth, I was pretty shocked at the stuff that came out because within a few minutes, I was able to build a local SEO machine that had not just built me a dashboard like this, like we can build dashboards in any platform, but this is actually a live dashboard that scans everything I need to, all the data and stays updated for this client. Plus, it was able to actually give me some real deliverables here. It was able to go and run several different modules of an SEO technical audit, a Google business profile audit, an on-page audit to give me this insanely detailed outcome, which would have taken me hours to do, to either set up an NA10 workflow or manage manually with different SEO tools. Even when I did a technical audit, wrote all the schema, markup code, everything that goes behind the website, it produced audit reports and that's amazing, right? Like to be able to audit a website just conversationally is cool. But what really blew my mind is it took that information and actually actioned the deliverables from this piece of software essentially that I built in Perplexity Computer and actually started from its audit Doing the work. So this is potentially weeks and months of SEO work for a client that it's just actioned automatically for me based off this automation. Speaker 2: So you're saying that you basically had to do like a full SEO audit on, sounds like maybe a client's website of yours. And then it didn't just stop there. Like it actually went into the backend. Again, I'm not super familiar with how SEO works, but like it went in and like implemented the findings of the audit. Is that what you're saying? Speaker 1: It did the audit. And this is a repeatable almost machine That's built into Perplexity Computer now for me to run the same thing every time. Did the audit and then wrote all of the deliverables, all the code, all the page updates, everything I needed. And then I've just found before we got on here that I can actually connect it to my WordPress and tell it Go and actually make those changes, please. So yeah, it can do the full sweep of work for us, which is wild. Speaker 2: Yeah, that is wild. Okay, awesome. Yeah, let's keep digging deeper. This is cool. Speaker 1: It literally like my heart stopped beating for about three seconds because I was like, wait, what? This was my whole job. This is insane. And I'll give you a little preview of how it sets up actually, shall I, before we get into like the behind the scenes of it. So now, anytime I want to run an SEO audit, for example, all I have to do is say, load my SEO Load my audit, load my local SEO audit, please. Or load my, sorry, let me try that again because I'm not quite getting the phrasing right there. Run my local SEO machine. See, I'm still not used to prompting this in just like loading up my machine. So we just dropped that into the perplexity computer box there. And now you see on my screen that it's actually loading up this Essentially, what is akin to a piece of software that I built, and it's asking me for just the key bits of information. So all I have to give it is, have you got a website, Corey? Speaker 2: Yeah, so you can use ReturnMyTime.com. And I know it's terrible from an SEO perspective, so I'm sure it's going to rip it to shreds. So this will be really beneficial for us. Speaker 1: Might take a little while to get the audit done, but we can jump into some other stuff. Do you have a Google business profile? Speaker 2: No, we don't. Speaker 1: No profile. That might not run so good. Do you have any particular keywords that you might think you might want to rank for? Speaker 2: Yeah, so like AI tools, Like AI audit, like we do, or like AI audit could be one and then AI assessment could be another one. And then something that we talked about implementing is doing a bunch of like blog posts or just like programmatic SEO around AI tools for XYZ profession. So maybe we could even say like AI tools for solopreneurs. Speaker 1: So I know that's pretty specific, but no, we love I love specific the easier to rank for specific stuff than white stuff. This might not work perfectly because it's for local SEO, but we'll see how we get on. Do you have any competitors like websites that you look at at all? Speaker 2: So, there's none that really come to mind. Of course, we have competitors and I'm sure there's plenty out there, but... Speaker 1: Shall we just throw mine in there for now just to... Speaker 2: Yeah, sure. Yeah, let's do that. Speaker 1: My AI one. So, these are the only bits of information that I put in here. And it's actually going to adapt the audit because it's got this AI intelligence built into it. So it's going to adapt its process a little bit. But that's all I have to give it and it's going to go off and run my AI audit now. I don't have to pay $200 a month for SEMrush or Ahrefs. I don't really know how to have to navigate an SEO tool. I just give it the information and it gives me back what I need, which is It's crazy when you actually zoom out and think about that's a whole person's week doing an audit for someone and creating these deliverables. That is the job. Speaker 2: And now that local SEO machine, so that's a skill that you wrote, right? Like that doesn't just come off the shelf with Perplexity Computer. Speaker 1: I initiated it and built it, but there are skills and underlying things that are off the shelf. I haven't fine-tuned this. All that stuff that I showed you coming out before was just what it thought was best to do. I could go and improve and tune what it can actually output for me. But it did a really good job just off the shelf doing these deliverables. I could teach it more tone and how I want it to write my pages based on the business and stuff like that. Sure. But you can see here, this is the key bit that I wanted to show you, is now we've got the three modules. So when you talk about skills in like Claude, you're like, oh, cute. It does one set of custom instructions. We've actually now got three modules running here. Speaker 2: So basically running three skills at a time in parallel, you're saying? Speaker 1: I haven't had enough time to investigate under the hood, but you can see it's just going way deeper than just running a quick skill. It's actually analyzing HTML elements on a On a web page, it's doing a competitive profile audit on other people's business profiles. It's doing an on-page analysis of all the text, all the pages on our website and comparing that to Google search results. And it's using different It's using Claude Sonnet for all of these, but it could be using Google Gemini if it shows that as a better source. So it's doing all of this at once. Whereas when you're doing something I think in like co-work, You see the plan box and it's doing step one, step two, step three, and it's kind of going things through one thing at a time. This is just doing three, not even three things, like 300 things at once under three modules there. Speaker 2: And now, so I know you said you haven't had time to get under the hood. And honestly, I haven't either, even though I've been using computer quite a bit. But do you know if there's a way to almost like export your skills from Claude and like import them into computer? Because the only thing stopping me from just using computer exclusively is that all of my skills currently live in like the Claude ecosystem. Do you know if there's an easy way to do that? Speaker 1: So for anyone who does not use a computer, you can hit this sidebar on the left here and go to skills, and you'll see all the skills in here. You can see my local FBO machine skill that's in there. So it looks like we can upload an MD file. So actually, we could try live now, couldn't we, and say, like. Speaker 2: Yeah, maybe while it's working, let's maybe look at that, because it's still working, right? Like, obviously, it takes some time. Speaker 1: So it's just running in the cloud, not on my computer or anything. So you can run, so with Cowork, you can run What do you call it? Scheduled tasks, right? But co-work has to actually be running and your desktop has to be on to actually make them work. With Perplexity Computer, it runs in the cloud, so you don't need to do any of that. What have we got here? Let's just print my brand. My brand skills, so this automatically applies branding to anything I output in Claude, all my logos and everything, which is a massive time saver and keeps things consistent across the company. So Perplexity Computer, to find your task or your chat, you go tasks and you can load up the conversation you're in. But we were in skills, we can hit create skill, upload a skill. See how we go. Now I've got to find where I downloaded it. Unknown Speaker: Let me just see where I downloaded that to, sorry. Desktop Elliot's Skill. Right, let's try that again. It doesn't look like it wants to... Speaker 2: Well, you can drag it straight from your Downloads folder in Chrome, right? Speaker 1: It's on the desktop, wasn't it? Elliot's Brand. Nope. I'm struggling to find it. Why am I struggling to find it? Speaker 2: You see what I mean? You can click the downloads at the top. Speaker 1: Oh, you know my downloads folder? Speaker 2: Yeah, just drag it right in. Yeah, there you go. Speaker 1: So only wants .md or zip files. So no, it doesn't look like it's gonna allow me to do that. Speaker 2: Yeah, that's what's frustrating about skills in general is that like I feel like nobody has really solved a way to just kind of have like a centralized skills repository where you can just kind of move them between all your different LLMs. Speaker 1: Yeah, I mean I would imagine we can just Go tell co-work to like can you turn this into a zip dot MD file or something? But it might take I with So the concept like I get the concept though, but yeah, I just was curious if that was possible But we could wrap it up in a in a different way to actually make this happen now Should I show you how I built this? This audit machine while it runs because it will take a little while to obviously do quite a lot of work there. Yeah Show people how easy it is to actually do this, which is why I haven't looked under the hood because I It just sort of happened by accident when I started playing with this stuff. You can see the way I started this conversation. I just said, what non-obvious SEO use cases are there for Perplexity Computer? What would feel magic and just not a research or a dashboard? Because a lot of people who I saw using Perplexity Computer, they were like, I built a dashboard. I was like, okay, cool. But like, what does this actually do? And it gave me a list of So you know search engine results monitoring. Content decay, analysis, technical audits, and a list of pretty cool ideas. I just said, can you build all of them for me? It went ahead and just built it for me. I didn't do anything special here except for maybe give it a bit of a steer. It went and just created this essentially deep skill bordering on a piece of software for me. We could probably go and show you, let's say like, let's go down here. We could take another of the ideas here, which was a topical SEO thing. So, you know, you talk about blog posts and stuff like that. You want to become an expert, the topic expert on AI audits, for example. This is what this idea would do. It would do all your keyword research for you, find every piece of information that you need to write content on for your SEO to actually become an expert and rank top for all your search terms. So I could say, I could take that idea and say, can you, can you turn this idea into our next SEO topical authority machine? And we should be able to run that task. And you can see I've got scheduled tasks in these machines that I've been building coming up to run every week, or as and when I want them for different clients as well, so I don't even need to think about half this stuff. Speaker 2: And so Perplexity Computer, because it's in the cloud, your computer doesn't need to be on for these scheduled tasks to run, right? Kind of like what you were saying earlier with co-work. It has the same ability, but your computer physically needs to be on because it only runs in your local machine. Is that kind of the difference? Speaker 1: Yeah, this is not using anything on my machine except for my web browser. Unknown Speaker: What are we talking about? Speaker 1: What's the keywords for this demo? I just want a repeatable system that I can then put any keyword into to start with. Do you want a one-off deploy tool or I want this package as a reusable machine or skill like our local SEO machine? And we can even say because we've got visual mapping and stuffing in here, we can actually get it to build out Anything visual, videos, video walkthroughs, all that sort of thing. So I've got my local SEO stuff as well, recording a transcribed and voiceover walkthrough of the SEO strategy. Usually in the past, I would have made a loom walking the client through things. I probably still will because that's touch. But if the client wants If I wanted to, I could send them an automated run-through or I could hand it off to my assistant VA and just give them a run-through audibly of what this all means basically. Speaker 2: And could you even do something where like you clone your voice with 11 labs and then it's like you record the actual screen share video of the tool once and then have an automation set up with 11 labs like, hey, anytime I get a new client, you're going to clone my voice and then demo this walkthrough using their name and their business name so that it is personalized, but obviously it's not you physically recording it. It's just the AI clone of your voice. Speaker 1: The other thing that blew my mind as well was just the sheer volume of connectors in here. Claude or ChatGPT, they may have like 50 connectors, would you say, something like that? Speaker 2: Yeah, I think maybe even like 150 or so inside Claude now kind of off the shelf. Yeah, this is crazy. Speaker 1: So I was building something before we got on this call. I was like, wouldn't it be cool if I could have a LinkedIn machine that goes through my messages, find pop leads, writes and draws replies. Nothing crazy that you maybe couldn't do in N8N, but it would take you a while to set up in N8N and get right. So I was just like, oh, I wonder if there's a LinkedIn connector. And there is. So it can hook straight into my LinkedIn. Yeah, that's what I said. Speaker 2: Because LinkedIn is like notoriously hard to get any sort of like automated access to. I figured that you were going to show this and be like, oh, and it's not possible. But this is crazy. I'm going to have to test this out myself then. Speaker 1: I can show you that little thing that I built in a second, but even now I just found another one. I'm like, oh, it's got this rapid URL indexer. So any new pages that it builds for my SEO website, it can then also link to this rapid URL indexer and I don't even have to go and submit and do any of that stuff. Do it. It will just submit it to Google and all the AI tools and the search engines. So there's so much in here that it's hard to comprehend what it can do, like being able to hook into LinkedIn and WordPress for me, Google Search Console, Google Analytics, all in one LLM automated procedure is I can't even describe what that means. Speaker 2: It's almost like it's more beneficial to figure out what it can't do, because I feel like it can do 98% of things. And then there's maybe like 2% of things where either like maybe it has the capability, but like the connector is not yet built in. And like, that's the only reason you can't do it. Like, is that fair to say? Speaker 1: Yeah. But I mean, you could get it to work with APIs. It could read API documentation. So this, I mean, those two processes look like they're still running because they're actually coding and doing stuff. Yeah, so it's still auditing your website, but that's a pretty hefty process. Speaker 2: Probably because we've got so many gaps, it's probably going to come back and just eviscerate us. Speaker 1: What I'm saying is like I'm a bit frustrated because I'm like, it's taken five minutes. Dozens of hours of work traditionally to do all of this or a lot of money paying a lot of SEO tools that expedite the process So this is the LinkedIn machine that I built literally just before we got on the call and I've masked some live data, but I gave it API information and stuff, which maybe I probably shouldn't do, but I just wanted to see if it worked. It was able to read all the documentation. It was actually coding stuff up in the background to build What I needed. So see my login token there. I'll have to reauthorize all these tokens now. Speaker 2: So tell me about this. Just like start with the actual connector itself. So like that connector that you showed us that can essentially connect to LinkedIn via API. What does that allow you to do? Does that allow you to scrape data in LinkedIn? Can you comment on what can you do? Speaker 1: So this is LinkUp API. So it can search, retrieve LinkedIn profiles and company information. So it can read, which is fine, like it can navigate profiles. It has read and write capabilities for LinkedIn automation, send messages and manage conversation threads. So yeah, exactly. So we can start doing all of this stuff in our Perplexity Computer workflows. Speaker 2: I don't think this was possible with LinkedIn. Speaker 1: Neither did I until I stumbled across this today. Be careful not to show my API keys again. Speaker 2: I'm going to add it to my perplexity right now as we're saying this. Speaker 1: Yes. This platform is like The API that gives it connection into LinkedIn. The connector runs into this system, which then links it into LinkedIn. And it does some stuff really well, like the documentation in here. Perplexity went in and read all the documentation about how this API works, so I didn't even need to Give it this it went and found it because it was struggling So it's got the LinkedIn engagement API that it's like okay messaging. How do I send messages and that went and found? All of the information in here to then build into our app that we were building. So like when I say I haven't looked under the hood, I haven't really had much to do except put connectors in place and then tell it what I want it to do. And it goes and is like building this whole This is a very interactive dashboard. Like I could see it coding the APIs together and all the different functions. And I said, you know, it can read my buyer signals. This is dummy data because I didn't want to share private messages on there. And then actually draft a reply for us. It feels way deeper than a Claude skill, like all of this is way deeper, but we can then pair it with those lighter skills that are my tone of voice or how I like to write, how I like to talk on LinkedIn. We can pair all of that to actually give it all of that fine tuning down the line, just conversationally through chat, giving it references, all that sort of thing. And then it's got built in the send via LinkedIn thing once we've addressed the, you know, tweaked the message or made sure we're happy with it. A link to comments about some recent comments that I've been having. And that'll scan every morning for me now and just update this like dashboard of people who might be interested in my services basically. And I don't have to like dig through messages and be like, who should I be contacting? What do I need to say to that person? You know, keeping track of comments. Speaker 2: So if I understand this correctly, like that LinkedIn lead machine, essentially web app that you built with Perplexity Computer, it sounds like you'd probably be way more efficient just managing all of your LinkedIn comments and DMs through that web app versus LinkedIn itself because it looks like that web app Essentially uses AI to identify which comments and which DMs are from real people, not just AI generated, and which ones actually look like warm or hot leads. So that dashboard even said like hot leads, there's two of them. So instead of logging into LinkedIn and being like, oh, I've got 15 DMs I need to respond to, you would just log into that dashboard and say like, oh, well, there's two hot leads. So obviously I'm going to respond to those first. And then it kind of helps you like prioritize, right? Speaker 1: It's more of like, I'll still go into LinkedIn and use it. But this is the place where I go when I'm like, what is the important stuff? What do I need to filter out from like, I've got 20 notifications. I don't want to see everything. I just want to see the core things that are going to drive I do revenue or good conversations in my business. That's what this is, like intelligence. And then I'm like, when I'm conversationally building, I'm like, can you make it reply? And see, I'm just testing, but I have an idea. Tell it the outcome I want. And at the moment, 80% of the time, if you've got the right connectors in place, it can issue this stuff and go and start building for you. So when I say you don't have to be technical to do any of this stuff, you really don't. And most business owners and professionals are probably still stuck, like 99.7% of the world is still on ChatGPT3 using it to draft emails. Speaker 2: Yeah. And I think that's the big opportunity for people like me and you is that like we have gone deeper into this stuff and we know that there's so much more it can do. And there's people, if they just knew what it could do, they would start using it. But a lot of people just kind of are limited by their imagination or limited by their current experience, right? Speaker 1: I think the blank canvas problem on these sorts of tools is really real. And you go in and you saw the first thing I did. I was like, what do you do? Like, what can you do? That's, you know, I don't know what these platforms can do because you see all the flashy advertising and you're like, okay, great. It's a new computer thing, Bob. So I went in and I just said, what stunned me, like create some magic. What can you do that nothing else can do? And it comes up with these. And then I tell it to go and do it. And every so often, well, pretty much every week now, there's a new platform that actually is delivering on the promise of what AI is in just conversationally putting your tasks out there and conversationally and It's doing all the work under the hood, and you're sat there as the ideas person, the strategist, the refiner, the human front end, and AI is taking care of everything underneath that's sapping energy from your day. You know, I've got to go through all these LinkedIn posts and comments and try and work out what's going on or whatever it is in your business. You can focus on what's the important stuff that actually needs my attention. Rather than getting bogged down by slack and everything and you don't get around to the work you actually want to be doing until like 2 p.m. in the afternoon and you're already exhausted and you're distracted, we can get rid of layers of this stuff in our businesses now. Speaker 2: And that's exactly what it's, I think, best used for. I mean, you pretty much summed it up perfectly. Now, I think we just got a ping that the SEO audit is done. Is that right? Speaker 1: And we've got two things built while we're doing this. I think it's run up our topical mapping system, but we've actually got the initial audit that we started with just tons of stuff in it. So we've got our Live War Room. Sometimes it does still error because this is new stuff when I try and load a full page app. So it's still like a bit hiccupy. But essentially, we've got your me versus you in our SEO war room. Speaker 2: I'm sure you're killing us. Speaker 1: I don't think either of us are doing too well here because none of us have got any keywords or those particular keywords ranked for. You've got way more content than I have. Apparently, I've got more social proof, depending on platform, I think. You've got, apparently, your 197 prompts beat my AI vault. And then it's got your untapped competitive edge. So, you know, you're not leveraging SEO or trust signals across some of these areas compared to other sites and competitors. You've got content gap analysis. So, AI audit, you don't have any position on here, but it's a high uncontested niche from its research. So, and AI tools for solopreneurs, high best ranking opportunity. So it was a good selection for you to actually go after. And then it's actually found some other long tail keywords for you to go after. You know, AI workflow audit, very low competition. Speaker 2: I'm surprised the competition is so low. Like the fact that even above it said like AI audit as a keyword is low competition. Speaker 1: I will tailor this in that I haven't given it a ton of feedback as to what... I haven't fine-tuned this. I only built it yesterday and started testing it. So I haven't actually... There are connectors that I could put in that are to dedicated SEO APIs that could actually pull up competitor data on this. So it is very much using just... It hasn't been trained on what makes a low or high competition keyword. You can see the potential of where this is going from a sloppy set of three prompts to be able to get this. It's pretty incredible. And then we've got like, the dashboard is cute. That's like what I can show to a client, you know, if I'm presenting back, it looks cool, like I can draw comparisons and stuff. But where the real stuff comes for me as a SEO specialist is stuff like this. This technical audit that is a boring spreadsheet, but it's where all the good stuff lies. And it's actually, it's technical audits I've been looking at and they're pretty bang on. Like some of the schema markup that's missing, like FAQ schema on the homepage, just some code behind it to bring out rich results in search results. Some stuff on the sitemap. All of these things are just fundamental things that need underpinning in your site and some meta titles are too long and things. These are actually pretty accurate because it's not relying on interpretation. These are key signals that are missing from your site. Because it's seeing all these, we've got the action deliverables. Again, I haven't trained this. I haven't got it. Sorry, I'll try and load that up again, see what we can see on the page here. I haven't trained it to my exact processes. But that's not to say I can't just go and get all my templates, a few of my YouTube videos about how I like to do local SEO, and I'd be able to train it in minutes, a transcript of my full local SEO process that I have on YouTube. I can just plug that transcript in and there's all my knowledge for it to work, to build, to fine tune off. But for example, it's gone and built you. Again, sorry. It's gone and built our AI tools for solopreneurs pillar page copy. So we can see it's actually gone and written our page copy with SEO intent in mind. So what our headers should be, where the CTAs should be, you know, other keywords, second header, what is an AI SEO, what's an AI tools audit, for example. Speaker 2: Again, this is so good, like, and it's, it's so clear that I mean, my business side, right, like, obviously, we have so much work to do on the SEO front. And there's so much low hanging fruit. But I mean, I would assume 99% of businesses are probably in the same boat that we are, especially like you said, local business. Like I know this is kind of your niche, like local businesses that have been in business for five, 10, 30 years and their website was built 15 years ago and they've never updated it. And I know, I mean, I know that's what you do. Like you just got flown out to Japan because you implemented some of these fixes for a You know, a company in relatively easy fixes and you blew them up like overnight. And this stuff is powerful, you know. Speaker 1: And I actually ran that. That was my first test for the guys that that company who are just crushing it now. In fairness, they had a fresh website for me to work on a new branch. So it was it was pretty fun to have a freshie, which is rare. And I ran this audit straight away. I was like, this is my First site I'm going to test because it's one of my top forms and it came up with some like pretty valid points of Trying to find it here for I don't know where it's gone That was in that one It came up with some really valid points about it read their Google business profile competitive direct to their best competitor and gave me the Like unbiased because I'm biased when I'm looking at this stuff. I was like, well, I've done the work So it must be pretty good It gave me the unbiased side-by-side snapshot of what was going on there. For me as an SEO person, it gave me the immediate next steps and the immediate next strategy when the owner is like, well, what do we do next? Sorry, I'm trying to find it here. But you saw what they came out looking like as. Speaker 2: I remember you showed it earlier. It was like a side-by-side, essentially almost like infographic-looking comparison of like, hey, here's your Google business page. Yeah, exactly. Speaker 1: This is actually my mom's old business that I use as a camera. It's closed now, but I used it again. I was like, let's run a side-by-side comparison of, again, SEO tools can do this, sort of, but they cost a lot of money. I know they just lack the flexibility to build them how you want. So you can see like, where are we losing here? We've not got many photos with our service areas are only some cramming nearby versus actually naming the exact towns we service. It's just like giving us a win and a lose. So now as me, I'm like, okay, all I need to do is make sure these are all wins. Speaker 2: You're going to dominate that competitor, right, I would assume, in the Google search results, if that's the case? Speaker 1: In theory, yeah. Speaker 2: Right. Speaker 1: It usually works. 90% of the time, that's how it works. Logarithms are out of my control. So, you know, it's even got, like, then I was like, oh, I can put Google Search Console in here? Hold up. Now I'm getting live data from my website into my LLM. I've got Google Analytics pointing in there as well. And as I said, in theory, I hooked up WordPress so that we can try live actually jumping around so many tasks here because I hooked up the WordPress.org connector, which I've never seen that in any LLM platform. So we could say, can you issue the changes direct to the website? You've got the WordPress.org connector set up for Smithbrook tuition. So because we have those deliverables running and in a document pre-written, the pages we need to create some of the meta titles, I'm hoping that actually We'll just be able to push that live, which I wouldn't suggest anyone does without checking this stuff and making sure they're happy with how this stuff is coming out. But this is a sort of... Speaker 2: It's a good exercise to know what's possible, right? It's like, if this is possible, which it looks like it is, then... Speaker 1: Because if it was, actually, let's get a page drafted up and written and SEO optimized. Let's get it drafted up and pump it into WordPress. It's like the bulk of it's done. Then we, the final step, we can go and do those final touches and be like, ah, let's just tweak those words, make that tone better, move that section, add this, you know, rather than starting from blank page. It is going to see if it can run the connector. But in theory, it should do because it is connected up. But this is the sort of stuff that I think People need to have with AI is not treat it as a tool and say, I tried to do one thing and it didn't work, bad tool, it doesn't work. The mindset often is take a washing machine, you get your dirty clothes, you put them in the washing machine, and if they come out dirty, you're like, that is a bad washing machine, it doesn't work. If you take that mindset to AI, You're not going to get very far because if something comes back and it's not right, you don't abandon the AI platform or the tool you're using. You need to treat it more like a person or an employee and say, give it feedback, give it some more Resources, more information to actually in the context. And so I like to say that AI is a whip smart intern. It's freshly graduated. It's got all the knowledge in the world, but it has no real world experience. So if you take a freshly graduated A grade student and put them in your business and just say, Go to work, they're probably not going to do the work the way you want it. It may be Wikipedia correct, but in a missile context of business and how it fits in with the rest of your team and your staff or whatever is running. That's how I like to approach AI. Refine, always improving, always getting better. Testing, seeing what it can and can't do. Disconnector might not work. It might only be able to do certain things. So here we go. So what can I do right now? Create a new blog post. So yeah, it can push new content on the correct URL structure into our website. So it can do some stuff. But it needs me to put in the passwords and stuff and the logins for the website in the connector, I think, to actually make that work. So we're pretty close. Speaker 2: Yeah. I mean, again, just a little bit of extra setup and it would be doable. And to your point about, like, I love that analogy about treating AI like an employee. I mean, it's no different. And usually if the output is bad, it's because the input was bad. Like AI can do Pretty much anything these days. And so if it's not doing what you want it to do, it's your fault. Either you didn't give it the correct information or you didn't give it enough information. So, yeah, I think you pretty much nailed it when you said that. Speaker 1: Yeah, it's. Once you start understanding how to move forward with this and how to build the base layers and the knowledge behind this stuff, things just exponentially grow. I've got all the information I need to actually turn this local SEO audit into a fully functioning thing that would do It's a good job rather than being like, it's pretty good. Actually, it does most of the work, but I still need to wade in a lot. I could get it there with another days. Here's all my frameworks. Here's all my knowledge. Here's all my content about my SEO business and how I like to communicate and the important things to me. Those can all be hard-coded into the rules of what's been delivered. Like, better than an employee because you have an employee and they go and do things the way they want to do it. Speaker 2: Right. Speaker 1: Give AI the rules. Speaker 2: Exactly. Now, and that's, this has been a great breakdown. I mean, are there any other kind of final tips or any other tricks you want to share when it comes to Perplexity Computer specifically? Speaker 1: Only in just some of the setup tips here of if we go back to our Let's start a new task here. You've got all the normal functions in here of dictation and voice mode, so you can talk back and forth with it to craft your ideas or even you could probably build some voice machine in there. I think you said about 11 labs. I'm sure there's a connector to actually get your own voice or your own AI avatars in there because it can generate video. The other button here, the plus button, It's got all the upload files and everything, add from cloud, connectors and sources. It has all those cool sources that Perplexity has, so you can search through just social, so you can just tune it to be like, I just want you to draw information from Reddit. Academic papers as well. But the real interesting thing and the big thing about Perplexity at the moment is you can switch between different models. So your main overarching orchestra, you're the conductor of the orchestra and you've got a sub-conductor which is your Claude Opus, GPT-5 or Claude Sonnet. So you can pick whichever one you want to be your lead AI and then it goes off and selects its other Out of 19 LLMs, which it wants to use underneath. So those are the two real big things to take away. It runs in the cloud, so you're not got any security problems of co-work or something similar running on your laptop and having to download files, which is a pain sometimes. Two, you're working with all of the LLMs at your fingertips, not just one. So you don't have to choose between, you've just got them all there. Those are the two key things I think to keep in mind. Speaker 2: I love it. And those are such good tips. And this is a fantastic breakdown, too, especially for anybody that is looking to get deeper into the SEO space. So, Eliot, before I let you go, where can people follow you or find out more about you or connect with you? Speaker 1: You can find me on YouTube, Eliot Prince. Just search that up and you'll find me talking about all these sorts of things and trying to just give helpful information on that. Yeah, you just search my name up on LinkedIn or Google, Eliot Prince, and you'll find the right place for me. I do try and be human front end. So my email, when I reply to emails, that's still me. I do reply me being LinkedIn. It is me on YouTube because I said I want human facing front end and I think AI is the under the hood, you know, driver cutting out all the boring stuff, surfacing the right stuff at the right time for you. Speaker 2: Absolutely. Yeah. And we'll have Eliot as a collaborator on this video too. So you guys can literally just scroll down if you're watching on YouTube and subscribe to him directly from this video. So Eliot, thank you so much for the time, man. This was awesome. And for everybody else that's watching, we'll be back soon. Speaker 1: So good. Thanks for having me, Corey.

This transcript page is part of the Billion Dollar Sellers Content Hub. Explore more content →

Stay Updated

Subscribe to our newsletter to receive updates on new insights and Amazon selling strategies.