How to Show Up in ChatGPT as a Local Business

Ryan Doser· 31 min· 5,833 words· 27 min read· English ·Watch on YouTube

This is the full transcript of How to Show Up in ChatGPT as a Local Business, published on YouTube by Ryan Doser. Every paragraph carries the moment it was spoken, so you can click any line to jump straight to that point in the video, search the whole thing for a word, or copy it out.

0:00And SEO has always been blackhead at its core, right? The whole game is trying to manipulate the Google search algorithm to rank high. It's not necessarily a search interaction. It's a conversational interaction. So, we need to be thinking about what sorts of conversations our potential buyers are having. I could log into chat GPT, search for something or prompt something. You could do the exact same thing, and the results

0:21would probably look way different. There's not a lot of similarities on a onetoone comparison of me prompting chat GPT versus you or anyone else prompting chat GPT. And what's really interesting is the self referencing list of calls are read by chat GPT but they're not referenced by ChatGpt unless there are third party validation articles offsite. If Chat GPT isn't naming your business when people ask it what to buy, like

0:44which websites is it trusting instead?

0:52Welcome back to another edition of AI rabbit holes. This is a podcast where we go down the endless amount of rabbit holes in artificial intelligence. And today I'm happy to have back on once again Chris Pantelli. Now Chris is the co-founder of Linkify which is a digital PR agency and also he came out with a recent study of a 21-day study of what chat GPT sites when people ask it what to buy. So that's the topic we're going to be diving into today is what is or how is chat GPT deciding who it sites in its results. So Chris, first of all, great to have you back once again. I had

1:25you on the podcast, I want to say over a year ago at this point, so it's good to see you again. Wow. Yeah, glad thank you so much for having me on Ryan. It's always a pleasure. Um excited to dive into this fastm moving space. Yeah, absolutely. So let's get right into it. Right. So, if chat GPT isn't naming your business when people ask it what to buy, like which websites is it

1:46trusting instead? You know what I mean? So, like let's just start from a high level of how do we how do someone actually get cited in chat GPT? Where should someone start? Yeah. Um, so yeah, it's it's it's a very nuanced thing I think AI vis visibility and how you're appearing in LLMs. Hence the reason that I've been conducting these studies. AI is a bit of a black box in terms of like the one side of it as a marketeteer, like how we want to appear in these things, but it's also a secret weapon when it comes to the

2:16ability to actually test stuff at pace. Um, so I've done we've been doing a bunch of studies. Um, and the one that we'll probably discuss in a bit more detail was a a 21-day study just because um, I thought it would be really interesting to to see um, if we ran the same 30 buyer prompts every day for 30 days consecutive consecutively, how would those results look on a you know dayto-day basis, week- toeek basis, like a you know, a longer period of time. Um, so we can dive into that. some interesting results there. But uh high level um I think you need to try and

2:51have an understanding of what it is that your ICPs, your customers, your potential customers are searching for. And I use the word search loosely there because don't forget the interaction that we have as a user with LLMs and the chat GPTs and perplexities. It's not necessarily a search interaction. It's a conversational um interaction. Um, so we need to be thinking about what sorts of conversations our potential buyers are having. Um, and then we need to look to see, you know, much like Google that didn't reveal how its algorithm worked, but you could see the end results by the first 10 first 10 results on its page

3:32one. Um, we have a look behind the magician's curtain when it comes to the LLMs. We can look at the sources and citations that it's uh that it's telling us that it's looked at. So optimizing your brand and your business for what the LLM are telling you it trusts and where it's looking and where it's going in order to analyze and spit out a result to a user's question is is the

3:55best place to start. Yeah, that makes sense. You know, and when you were on the show last year, you said Chad GPT sources were full of tier one media. So this makes a lot of sense to me, but first let's define what tier one media is for the audience who hasn't heard that term or understand what that means. And do you see this as a similar trend today? Fast forward to 2026. Is it still full of tier one media? Are we starting to see more tier 2 social media, maybe some forums like Reddit in the mix? What are you seeing from today

4:24versus what we saw a year ago? Um, okay. So yeah, tier one media was and I define tier one media as as uh large journalistic uh authority websites um that have a a trusted history and they produce content with a with like a trusted monetization model. So um they'll often times be household named brands like you'll recognize them CNBC, Bloomberg, BBC. That's the creme de creme of tier one. And then obviously you have a bunch of you know news sites um and authority uh content sites that you will have also heard of. Um and they're producing content in a journalistic or editorial uh ethical

5:08way. uh they're often monetized as well. Uh but that monetization path is clear and obvious and and disclaimerred and blah blah blah. Um we are seeing now the dominant sources across many categories being uh like listicles, tier one media um and then um also uh like review directory pages and directory sites. So again it varies from industry to industry and this is actually super interesting how much it does vary from industry to industry and how much um sort of the the the the difference that you get the volatility within those

5:55industries is quite significant as well. So certain industries um maybe you might consider it to be a more evergreen industry or or a slower paced industry. The LLMs will leverage its like training data. it will do less grounded searches and it will um produce results that don't vary as much dayto-day session by session whereas more fastm moving and um industries that are are more competitive maybe um the results can change significantly and you will see this you only have to go into chat GPT and in incognito mode turn off uh the memory run a search do it an hour later do it a day later at one minute you could be Oh

6:38yeah, great. You know, Linkifi is in for best digital PR agencies and 20 minutes later you're not in there. So that's pretty interesting. And then what's what else is also really interesting as well is the the amount that chat GPT is leveraging um that personalization. Um so this is something which is extremely difficult to optimize for like you can't optimize for individualistically optimized searches housed inside of of these platforms. Which is why when you're testing uh a really important distinction is you don't want to be testing against the API which will give you API results which are very different to the in browser results experience. Um

7:19so not overly complicated just you know you can easily spin up a test connect to an API but those results aren't going to be sort of um generic like you want to test let's say like a thousand random people in the US are searching for best med spa to get a nose job. Let's say if you do that against the API, the results will be wildly different to um if you were able to run them uh in browser a thousand times. So that's an important

7:42distinction as well. And that makes sense. And maybe this might be a good transition to screen share of that study that you ran. And maybe we can actually look at some of the results. And while you're doing that and transitioning over, Chris, one thing I always ask a lot of people in this space, you mentioned listicles. Okay. So, there's been a lot of studies with ClickUp and I think there were some other ones that ran these studies about self-referencing listicles because obviously it makes sense from my website or your website to position yourself as the number one digital PR agency and you're going to spin up a million

8:11listicles trying to manipulate these LLMs to site that. Can that actually hurt you in the long run when we're doing these self-referencing listicles without a bunch of third-party mentions and we're just doing it from our own content? Is that am I kind of making sense with that question there? Yeah. And Google has clapped a bunch of sites that have completely um abused this. So like don't forget when we talk about like optimizing for Google organic SEO and then optimizing for AI like Google is an AI now as well with it AI overviews and it AI mode. So if Google is going to clap you on its SERs for its

8:47organic and then it's also going to like clap you on where it's going to position you in in its AI. Don't forget AI mode is now like basically 50% or more of queries or basically it's going to be default mode no matter what. So yeah, that uh that and that self-referencing um thing is we've recently just done another study um across real estate brokerage owners um inside of the US um and these listicles and what's really interesting is the self-reference like referencing um listicles are read by Chat GPT um but they're not referenced by CHPT unless there are third party validation articles offsite um and then

9:27they also like to see a consensus of that information on those third party validation sites as well. So, not just your website says you're the best and then you've done like a thousand press releases that saying it's the best that is obviously a press release. Um, but if it's realtor.com and architectural digest saying that you know you're the best or you're the second best and you've put on your own site um then you'll get a consensus. They'll read it and then they will output it. So, it's like the old days of SEO black hat like how far do you want to push it? what's working now almost certainly will be a

10:00different landscape and when you have me on in a year's time Ryan for our yearly podcast like I don't know but at the minute we can only derive from the results which are in front of us yeah so why don't we dive into the study here the AI visibility volatility index was there any like kind of unique data points that you guys found I guess what are like the top things that stood out here I know this is a long probably PDF and we don't need to go through this for the audience but what are just top one or two things that stood out to you in

10:26this study Yeah. So I think the important thing is just quickly how we set up the study. So we took 30 buying questions every day for 3 weeks and measured how much it's recommended sources changed. So the sources which if you go into chatpt and you type a query at the bottom um the sources is different to the citation. So a citation is where the business is

10:47mentioned within the output of the LLM. Um, often it will look like, you know, maybe a backlink, a contextual backlink inside of the text or it'll actually reference it, um, you know, within that text. The sources, it literally says sources and these are the the documents that it's gone and looked at from its query fan out. So, it's it's fanned out a query and it's gone and looked at these particular um, documents. So what what we wanted to see was not necessarily the you know the brands that were were were being repeated over over these 30 buyer questions run over 21 days but how much the actual sources changed because

11:24if you're going to optimize for certain sources then it's good to know that you know if you're going to spend budget and time optimizing for Reddit and actually Reddit was all the rave last week and now it does not not all the rave this week. Is it worth your effort? So having an understanding of the sources and how the the the actual sources change over a period of you know a sustained period of time. Um so yeah I've got a nice one minute version here. So Chachi PT's recommendations look chaotic daytoday but underneath they are surprisingly steady. So this actually is probably not that surprising. um as you would expect

12:04if it if it was completely chaotic every day it you know it it indicates it doesn't really know what it's doing. Um so having like a relatively steady um set of sources means that it's uh it's trusting certain places and it's going back to those same places again. Um so if you ask chatpt who the best personal injury lawyers in Phoenix today and again tomorrow the list of websites it sites will look mostly different. The study shows that almost all of that difference is randomness. The same thing happens if you ask twice in one afternoon. Once the randomness is measured and removed, real change is slow. So,

12:41can I stop here, Chris, for a second because there's a good distinction here because we're talking about local results, local citations in chat GPT with that Arizona example versus a national citation result with chat GPT, right? So when you're doing local, that makes a lot of sense that the volatility is not going to be as high as when we're doing like a national type of query or prompt where the volatility is a lot higher. And so like in my head, if I'm trying to invest in like an AI visibility or AI search visibility tool, I think a lot of this is still up in the air pie in the sky in terms of national

13:14on actually understanding like what shows up. Because where I'm going with this is I could log into chat GPT, search for something or prompt something with no context or I have context or I'm on a different computer or a different setup. You could do the exact same thing and the results would probably look way different. There's not a lot of similarities on a onetoone comparison of me prompting chat GPT versus you or anyone else prompting chat GPT because it's never the exact same prompt with the exact same harness. So like where am I? Does that make sense where I'm trying

13:45to go with this? Yeah. No, no. Yeah, exactly. And this was all um local uh that we did. Um I I I proposition that local might be more volatile than like national stuff just because of the like the training. If you search best CRM tool, like it's always going to the same sources. You're always getting like HubSpot and like the same results being outputed from like these well-ingrained and trusted like default trusted listical sources. So that was my theory. Um so we started with this. Um and as well like these are likely less competitive and easier to manipulate if you have the time and the budget. So,

14:29for example, if like you it's very very hard no matter what the source is for best CRM, it's going to be a lot more effort, budget, and time to get yourself to be outputed as best CRM. If you're like a brand new CRM tool, let's say it's going to be very, very, very difficult. Whereas on a local level um if we're seeing that actually catch PT is time and time again go to like a certain uh industry listical or a certain um directory let's say um then a lot of times these are things that like local businesses can go and claim and set up but haven't done and therefore

15:07the ROI on that effort is going to be like instant and significant. So I think that was the reason that we well the reason that we conducted the study on the local level but yeah I agree with you um the difference between results across users can be substantial and and the only way to sort of get accurate results for that is by doing the test in the way that we did which was non-API um incognito browser testing. So that theoretically could be any user anywhere that doesn't have any saved

15:41um sort of memory. Does that make sense? Yeah. So someone listening to this, I mean tactically speaking here, Chris, someone listening to this and they want to get cited in Chad GBT, let's just say they're a local business. We'll just throw that out there because this study was local. Obviously, first thing they should probably be doing is going after these local directory sites. You know, Better Business Bureau, Yelp, you know, all these sort of local directory listicles. Maybe a subreddit, a local subreddit might be a good play. And then also creating local pages on their website to try to get that trifecta there going is if there's something else

16:12I'm missing, probably press, you know, get digital PR, uh, press releases, getting covered in the local media is huge for this obviously, but like if someone's listening or watching this and they want to figure out, Ryan or Chris, how can I get cited in chat GPT consistently from a local level? What would be your advice to them based on

16:29the results from the study? Yeah. Well, first of all, you need to like run your prompts and derive a baseline of the sources that are that are there. Then you need to go and get the the the sources that you don't have. And a lot of times we're seeing especially on a local level that chat GPT isn't necessarily recommending individual providers, facilities or businesses, but actually it's recommending the people behind those businesses. So, if I'm having a conversation with Chachi and I'm looking for uh real estate agent broker, real estate agent in um Miami, Florida to help me with a firsttime home buying or I want to buy a a house with stables,

17:10let's say, because I've got horses. Um it's not necessarily or or more often than not is recommending individual brokers from firms as opposed to firms, if that makes sense. We're seeing that across the board. So, if I'm looking for financial advice, if I'm looking for medical advice, let's say I need a procedure and I'm interacting with chatt and we're doing those uh you know buyer queries, uh I've recently had um I don't know, let's say I've recently had a nose operation, it didn't go very well. I'm looking to for you know, for a second operation to to help fix it. Uh I'm in the the Miami area. Like, what should I

17:46do? It's not nec it's not coming back with you want to go to Miami nose repair center because it's the best you know it doesn't do that. It says you want to go to Dr. Joe Blogs who is you know one of the area's leading specialists in nose repair surgery and he works at or his business is so you want to be first of all having your business and your representatives of the of the business as defined entities. You want to make sure that your business is in all of the places, all of the sources that you can see chat GPT is trusting already. And then you need to build demonstrable

18:22trust and authority to the person within the business because chat GPT is looking at sources, okay? It's making its decision on whether or not to trust the information from those sources based on the wider trust and authority signals that it can see. And and we just know that. I mean, we've been working with a real estate broker in Texas, Austin, Austin, Texas for 5 years, and you go into Chat PT now and say, "Who's the best?" His name is coming up time and time again. We've had him there as a talking head on CNBC and Bloomberg by

18:52lined articles in the New York Times. Like Chat GPT and the LLMs, they see this, they this is how they can validate and and trust like like Google's genius behind Google's algorithm was how could it rank a page that it only can understand through m through machine learning. You know, it's an algorithm. is not a human. How can it know whether or not what page to put what to put on spot one for, you know, can I cure my cancer? Like that's a deeply important topic and you could see that that is ripe for, you know, manipulation and abuse. So, how could Google, you know, hands-free algorithmically know which

19:29pages it can trust to put on spot one, spot two, spot three? It needs to measure trust and authority. And it did that genius, you know, genius way through its backlink uh page rank algorithm. And we're seeing the LLM having to fight a similar battle. And the way that they're doing it obviously is they're able to ingest a huge amount of information. Um, and the information that it can deem a more trustworthy um is the information that it then validates against, you know, things like listicles and stuff like that. So, it's

19:56Yeah, it's exciting, fast-paced. Yeah, that makes sense. You know, and SEO has always been blackhead at its core, right? I think a lot of people want to conveniently ignore that. There's the whole game is trying to manipulate the Google search algorithm to rank high. It's no different with GEO, AI search, AI SEO, whatever you want to call it. Now, we're just shifting to AI search and LLMs. The game is very similar. We're just shifting different mediums. And Chris, one other question I had, you know, based on this, we're talking a lot about chat GPT, okay? But from what I've seen, the data that I've seen, the people that I've

20:26talked to, most search volume still happens on Google. I'm not talking about traditional Google results. Maybe that's Google AI overviews, Gemini, Google AI mode. So, does a lot of this apply still to Google's, you know, in their AI properties and same with Claude and Grock and other LLMs. Are a lot of these trends still the same across other LLMs

20:45and AI search engines? Great question. Uh, I I will be honest. I think I don't know and I think anyone that says they know that I don't know how you could possibly know. like we're focusing on two forms of LLM visibility as an agency as as we do social content as we do our research and studies. So Google uh you know organic search results and Google AI overviews is housed in the Google section and then we're looking at chat GPT as that sort of biggest user adoption for the wider market. Um, now I use all the tools I, you know, I use Claude exclusively for coding. Just

21:27launched the rebuild of link website today actually. Um, all of the internal tools we build we're doing through Claude. Um, we're using Grockbot for a load of automation stuff. We're using um, which is X's, um, X's AI. Um, so but in terms of actual like running these tests and and and what what we think is going to be the I guess as the biggest uptake in terms of like user adoption than Chachi BT but to speak to your point yeah I mean Google is still by far the biggest use case when it comes to AI and as it shifts towards full-blown AI mode it will it will you know it's going

22:09to be winning that that competition. Um, so I guess really to read between the lines, your question is maybe do we optimize differently for Google organic than we do for like Google AI overviews? And then why would we be doing a different strategy for that if we then specifically opt in for chat GPT? Hey, look, I own an agency where our core goal is to build your brand, build trust. That works anywhere. like Nike is going to get recommended more for like best trainers than you know Ryan Dosa trainer brand that you just launched last week because nobody trusts you, nobody knows you, you're not a brand. No

22:49offense, okay? I bet they're awesome trainers. Um, Nike has spent decades building demonstrable brand trust and authority. So the overarching play, the the best thing that you can do above all else, you know, listical optimization, you know, Reddit thread injection, whatever's working today, whatever's working tomorrow, it worked last week, it's not going to work this week. Forget all of that. The main thing that you can do is become a real brand, uh, and build real trust and authority for your business as an entity and then for your your expert representatives, the spokespeople of that business. That will have longevity, that will survive into the future. Um, but you know, clients,

23:28businesses, they want to be able to see results and you know, they're checking and and quite rightly, um, if you had a local business that was getting 60% of your, you know, referrals from the map pack or from direct calls from, um, you know, those Google features and that's declined. You need to improve

23:48that because your business is suffering. So we do need to look at those sort of uh as you said it's always been a manipulation game and that is what we need to look at in the short to medium term alongside those longerterm authority building uh activities. Yeah, I I like what you said there and that might be one of the concluding points we can have of this episode if you want to bring it back here is just shortterm to midterm focused on that, you know, manipulation, getting in the best of directories and doing those sort of things and then the long term it's all about building a brand, building

24:18authority both on the personal brand side and obviously of course on the business brand side. Now, one of the final questions I had for you, Chris, is that g get your crystal ball out. Let's say 12 months from now. What would you say the number one way AI changes which websites are getting trusted as sources

24:35that nobody's talking about yet? Uh, it's got to be agent ready, I think, hasn't it? With the launch of Meta's um Metam Muse. Yeah. And then, didn't Grockbot just released? So, agent ready. I don't know if you've seen this, but Cloudflare actually just put out a tool called it's like an AI agent readiness tool. It's similar to Google's page speed insights core web vital metrics where you can actually plug it in and Cloudflare, you know, puts all these things about robots.txt,

25:03LLM.txt. There's this thing called webmcp, which is kind of a new protocol now, right? So, look into this. I'll leave a link to it in the description. I think you're right. I think this is where we're going where if you're neck andneck with a business and you have branding and you have authority, well then there's all these other things that you have to optimize for to make sure that your site is quote unquote agentic ready is kind

25:23of how cloud play defines it. Well, I mean that's exactly I mean it has to be that it's going to be I mean I've started interacting with chatt now on the I don't know if you've pressed the little yellow button now and the voice mode the the voice mode, right? So I'm driving my car, right? Let's say I get in the car and I've had an idea for a new funnel or, you know, and I'll I'll whack this yellow voice mode thing on and I'm having a full-blown conversation with what feels like a human. Um, so Facebook's Met Amuse and whatever the other aentic ready sort of consumerf

25:58facing um personal assistant I think is the right. Yeah. Yeah. And it's going to just be you know the wasn't the adoption of of Metamuse on the because I've tried to access it. It's not ready yet in the UK, but wasn't it the most downloaded app within a matter of hours or something on the app store? Because people trust Meta, they trust Facebook. So maybe we haven't seen widespread adoption of agents and agentic workflows. And why would we? We're like, you know, kneedeep in in this as is as as people in the AI space, but you know, mom and dad and nan and granddad, they're not necessarily

26:32going to be using this and they might not even fully understand it. But when this sort of beautiful housed thing inside of Facebook next to stuff that they know, groups and marketplace and oh now do you want to buy this? Yeah. Okay, just add your credit card, we'll buy it and it'll get delivered. So any website that is set up to function in that way is going to have more favorable uh interactions and and traffic and all of that from whatever these agents are compared to let's say a site that has accidentally switched on block robots

27:05access in Cloudflare which you can do. You know you can block the robots. So don't block the robots. No that that all makes sense. Um Chris, any other final thoughts here as we just conclude this conversation on, you know, your chat GPT visibility study you did? Um, you know, where this authority and visibility, search visibility is headed going forward. Like what would just be your final advice to the audience here who maybe they run a business, maybe they have a personal brand, maybe they have a YouTube channel or or something where they're just trying to continually figure out how can I show up and get cited consistently in chat GPT AI search

27:37and other LLMs. What's your final piece of advice here that's actionable? Um, go and interact with with your LLMs. First of all, just find out if you are getting um any sort of sizable attribution from any platforms currently. Like I know we work with a few clients in the finance space and X is like is a much more significant sender of uh of traffic than um than chat GPT. So just first of all get a sense if there is like a platform that's

28:10winning out within your particular area. for mo for the most part it's going to be chatpt but there may be nuances once you have that idea of like okay so I think you know people are going let's say you're a local law firm and you're a personal injury law firm people are inside of chatpt and they're saying you know I got run over the other day my legs broke you know what what can I do where what's the next steps where should I go um start asking some of these uh buyer intent queries inside of CHP see who it's recommending which one of your competitors is recommending and if it's

28:41not recommending ing you ask chat PT why did you not recommend my law firm you know name your law firm and Chachi PT will come back to you yeah it will say I know you um but this firm was in uh three trusted directories they claimed their profile on X Y and Zed um and the the lead lawyer at that firm is certified in X Y and Zed and uh has been quoted in a number of you know peer-reviewed peer-reviewed articles blah blah blah blah blah So, it's going to give you a pretty decent road map and blueprint to go and achieve the same results. And then test as well. Test,

29:18test, test. Periodically be testing because this stuff changes all the time. Yeah, Chris, that's really good advice. And, you know, I like your point about reverse engineering what's already showing up. Most people don't even think about that. They just want to know what's the latest trick and what can I do. It's like go actually search, see what's showing up and figure out how you can reverse engineer to get your place

29:36cited in whatever is showing up. Right? That's very simple advice. But with all of that being said, Chris, thanks again for hopping on. Where can people follow you online to learn more about what you're doing? And do you have any any other studies or anything cool that you got in the hopper that we can look forward to here in the future? Uh yeah, go to linkify.io, follow me on x um

29:57linkify_. I've got new studies dropping all the time. Um I'll post them on social, so if you just follow on social um and then I'll be putting them on the website as well. But um yeah, loads of studies in the pipeline. Uh again, it's never been easier to put these things together. If you've got time, go and look at connecting some things like uh bright data uh data for SEO, um any decent tool that's got an MCP and you can you can put some incredible stuff together. So

30:24yeah, exciting. That's a whole other rabbit hole. We could get into how creating your studies is good link bait and you can get good authority and like get good press, right? That's a whole other topic. But anyways, Chris, thank you again for your time and thanks again to everyone for watching or listening to this edition of the AI Rabbit Holes podcast. Be sure to

30:39tune in again next time. Thanks, Ryan.

Where these words come from. This is the caption track YouTube holds for this video, written automatically by YouTube rather than by the creator. We read it, tidied the line breaks and laid it out so it can be read. The plain text version is at https://viewrankai.com/tools/youtube-transcript/cWxecZ74yr0.txt.

All rights in this video belong to Ryan Doser. Watch it on YouTube. If this is your video and you would rather this page did not exist, tell us and we will remove it.