Why AI Search Isn’t Recommending Your Brand

Full Transcript

0:00You cannot just rely on one platform anymore if you're not investing in AI visibility now. You're not a part of the conversation. Hey everybody and welcome back to Jolly's TL;DR. We are your AI search agency and the main reason we are your AI search agency is because of this lady right here, Muskaan Rana, and I'm so grateful you're here. Muskaan heads up our well we call her our director of operations but really you're the head of our R&D pioneering all of our AI search services. So without further ado I'd love it if you introduce yourself real quick and then we'll dive right in.

0:47Hey everyone, thank you for the introduction Greg. Um as Greg said titlewise I am the director of operations here at Jolly Search but last few months have been really really fun. Uh I've been building various products that would help our AI search service and yeah before Jolly I was a project manager then I turned into an operations manager and now I'm here. So yeah I've been uh in marketing for majority of my career I've uh got like eight years of experience but out of those eight years five and a half years was spent doing work for marketing agencies.

1:31Um, and this is my first time working with a linkbuilding agency and expanding that agency. Um, to AI visibility. Uh, and I think it's been really fun. AI visibility is very fast, dynamic, still very nascent. So, it's been a fun couple of months lately. Yeah. You say fun. We We had to delay recording a couple times so could laugh uh get our laughs up. Uh but but in all seriousness, I'm glad you brought up your marketing background. I was saying last year it's been interesting observing SEOs realize that we are part of the marketing department. GEO, AEO, AI, search, whatever you like to call it, has kind of helped us all realize uh wait a minute, we're not just technical uh you know schema kind of manipulators.

2:14We we are actually part of the brand. And at that time a year ago, I put a quote in David Bain's book here, SEO in 2026, and I said, uh, SEO best practices, GEO best practices, there's like a 90 plus% overlap, but it's been a year. I I' I'd love to know from you because you've done the most research inside our team. How would you characterize the is are we still at like almost one to one overlap over the last 6 to 12 months or have we been diverging uh and to what extent? I think we still have an overlap but search has evolved just so much over the last couple of years.

3:04Um that yes there are certain practices that we were not doing as part of our SEO services that we now have to do for getting good AI visibility results. So yes uh your basic stuff is still very much important. um backlinks is still very much important. You've got your onpage, your technical SEO, all of these things are still very important. These are your fundamentals. And to give an example of a client recently, we were working with a client who just wanted to get started on AI visibility services, but when we did a technical audit of their website, it was looking very bad.

3:52So this is why good SEO is still very much important because there's nothing AI visibility experts can do to get you recommended if your SEO fundamentals are not quite right. So you got to first get the basics right and then you can only scale to AI visibility services. Okay. Uh we're we're going to get into best practices, bad practices, but I wanted to ask you you said the word recommended and that's been a big learning experience for us. I wrote about it a couple newsletters ago. It was number 89. Uh if you want check it out on the site, but can you for us right now Muskaan explain the difference between cited and recommended? Yep.

4:37I think this is a great question because um I do my research and I see a lot of AI visibility experts like even they are confusing this metric. This is a very important success metric for the AI visibility campaign. Um and these are two different things. So getting cited means that for a specific query on Google in the AI overview results tab you see the answer there and on the source panel that is usually either on the right hand side or it's on the bottom you see a list of websites that the bots went to and scraped and crawled to give you that answer. So if you're a source website within that list you become a cited source.

5:23However, if for that specific query, you're a part of the main answer body, then you become a recommended uh source for that query. To give you an example, we have a client right now who wanted to focus on a very informational query like um what is debt collection? Um, and if we actually were to go ahead and target that query, I think it's great if they were doing just onsite with us because it's good SEO content. It's a fundamental that you got to expand on to build your topical authority. Uh, but their goal was AI visibility and get recommended for their target prompt.

6:09So um instead we suggested them to have a prompt which is something like what is the best debt collection agency for marketing agencies. For that specific prompt our client actually gets recommended today and the reason is prompt selection. I think that is also very very important. Um the type of prompt you're choosing is it going to trigger brand recommendation? Is it uh going to trigger like you as a cited source? So, this always happens. Yeah. So, we've gotten uh I think Google's thumbs up to my answer. So, I'm going to shut up now. Yeah. As long as Google says we're on the right track, then we're good.

6:53But uh you know you and I were in Shenzhen in June of 26 and you gave a 4-hour workshop and you were walking them through the stuff that I overlooked which is part of this. You have to start with ICP. You have to start with identifying your persona. You have to think about their customer journey. And then you walked through these prompt cluster categories. And that's something I hadn't thought about and and so cited versus recommended the really simplistic way I've distilled what you've been teaching us every other week in our workshops is cited is an informational and recommended is a bottom of funnel.

7:32And I think when our clients or maybe other providers get confused about this is they're they're trying to get recommended for prompts that there are no recommendations. It's an informational prompt. And yeah, it's it's a it's a big learning curve for us. So I guess I guess with that said, you know, Jolly, we try to be as transparent as we can. Maybe not going to give the full recipe to the secret sauce, but what are maybe one or two best practices for those who want to get recommended, not just cited, not just have the bot scrape your site and use your information and put you over there as a resource, but who want to be the answer to who is the best at X.

8:12Yeah, I think if you want to be recommended for a prompt, not only for 24 to 48 hours, but you want to own that prompt, um, we have a service called Own A Prompt. And now we're just end of episode. Cool. You can buy our service and own that prompt. And uh now we're actually uh moving towards calling it own your category because we understand uh after working on this service so much that the things that we're doing to help our clients own a prompt, they're not going to end they're not going to end up owning just one single prompt. They're going to actually own multiple prompts that are just variations of that prompt.

8:53So the prompt that you're selecting needs to be very very important to you. Um, and it actually needs to cater to your target audience. It cannot be an informational query because that is like somebody who is not probably looking to buy your services. So, it's a bottom-of-funnel prompt that you're looking to target that will serve your exact customer, your exact audience who can afford your services and actually convert. Um, I'm going to pick a bone here. Um, I should say it or not. Maybe I should just say it. We can cut it out if you feel like that's too offensive.

9:33Um, I was recently doing some research because we have so many campaigns going on and it's so interesting to work with different clients because we have a client in the construction industry. they come up to us with different questions and the client who is an owner of a B2B SaaS comes to us with different questions and it's really interesting to answer all these questions because this uh service is not only new to them it's new to the entire industry right so sometimes to answer their questions even though I have the answer I have to do a little research to kind of present it in a way to them that will make sense to them.

10:14So fortunately unfortunately I have to compare this to the old SEO days because that is their understanding of uh search. So one of our clients actually pointed out recently that um they wanted to target um a category that is not their exact category but an overlapping category. So, this is something SEOs have done uh for a long long time, and I think it's a really nice strategy if you're trying to steal your competitor's traffic, but it's not a nice strategy if you're trying to steal traffic from a category that you have an overlap with, but they are not your target audience. Um, I'm not going to use the exact example here.

11:00Uh, I'll give you one. Uh, we we had a great client. uh this was a few years ago and the reason they came to us is they worked with a prominent SEO uh and they said you know they made a an awesome graph for their own socials out of us. We had a huge traffic lift but all the blogs were about you know national pancake day and about cat care and all these things. Well the client is B2B lending. That's what they do. They they don't actually have anything to do with pancakes and cats, but they did have a lot of traffic about pancakes and cats.

11:43So, it it's like you said, it's one of the main I think Yeah. reasons that SEOs have caught a bad reputation is we have fixated on metrics that I mean thankfully double-edged sword here, right, Muskaan? But thankfully traffic is just fading away as a metric anyhow. But when it was the primary metric, it ourselves fell victim to it, right? Evaluating sites based on traffic solely. So I totally get it. But yeah, I'm right right there with you. Uh we have have to think about what do we want to be recommended for. Yeah. Yeah.

12:27I think it's a metric that was built by SEOs to track progress and it is also a metric that could easily help SEOs communicate the success of their campaigns and I don't blame them. I think SEO can be really really hard and uh it takes about 6 months to see the actual results of the campaign that that the client has been paying for. Um so 6 months any client would get a little antsy and ask you hey what are we paying you so much for? So this metric was built that hey you see there's an increase in the organic traffic on your website but at the end of the day um everyone has their style of running campaigns. I think we want to bring the traffic to the client's website that is uh highly qualified.

13:07Um this is a term that they use in PPC. Usually it's marketing qualified lead. So you got to get them leads that are actually qualified to convert, not only leads that can, you know, not even afford their services or are not looking for what they're offering. So, I think starting with the right prompt is very important for somebody to get recommended and not go around with your old SEO tactics because it's not helping you get recommended now. And if you're not recommended for your category query, it basically means you are invisible to your clients.

13:53uh we have seen SEOs giving everyone a lot of push back over the past few months and I don't blame them. I think it was very nascent to everyone and everyone has opinions. While some people were saying that 10 blue links are dead, there were some people actually going against it that hey it cannot be dead. But here we are. Uh it looks like AI search is not going away and those 10 blue links are kind of dead. So you got to be in the AI overview like in the answer where AI is only recommended I guess four or five brands at max and not 10 different competitors.

14:35If you're not investing in AI visibility now, you're not a part of the conversation uh and the conversation is barely just considering a short list of candidates um that your user wants to evaluate before they make a purchase. So, not being a part of that conversation means you're basically invisible to your target audience. So, uh, picking a right prompt that your audience will be searching for is very important to get recommended. Um, I know you asked for two best practices, but I'm going to give like two more. So, I'm going to give three because this was fundamentals. So, yeah, if you've picked the right prompt, then I think uh you got to produce content that's got uh genuine data in it.

15:18Um, and when I say genuine data, I really don't mean that you can go on the internet and find uh somebody else's data and just use that data. Again, AI search is now prioritizing uh websites that are producing content uh that is actually new, genuine, fresh, and unique that nobody has read before. And the logic for that is pretty simple. uh these models need new content that they are unaware of to train them. So if you actually favor them that you're providing, you know, it's like give and take. You're giving them new information that they don't have and is valuable for them to produce their results.

16:03they're going to benefit you in return and rank you higher in traditional Google search but also they're going to register you as a credible source um in case somebody like your users asking uh a prompt to the LLMs um and you will get uh preferred when the results are actually generated. So you become a preferred source by them by producing that kind of content. And of course the kind of piece you're writing about. Um so let's say if we were writing a piece at Jolly SEO about hey we've run 250 AI visibility campaigns and here are top five things everyone is doing wrong. Uh and we will use our data that only we have. Nobody has that data.

16:52So they cannot make those data-backed claims or they cannot share those learnings. So that is only unique to us. So when somebody is searching about AI visibility in general, we will become a credible source for AI to you know basically go to and crawl for their answer. But also if somebody's looking for hey uh what is the best AI search agency out there then also we will be a recommended source because we have uh established our authority over that niche if we keep producing content that is um new and genuine and not copied by others. So I think having your unique data and I think every company has unique data that is just unique to themselves.

17:39Um we were talking about this uh I think a couple of days ago uh at our leadership meeting that we think uh this whole thing is going to take a full circle and we're going to become Jolly Content again. And the reason I say that is because I actually see us moving in that direction. Um and so for example, if somebody asks uh what is a debt collection agency like an informational prompt or what is linkbuilding? I think AI can do a better job at writing an article. And this is just my opinion. Um I think AI will do a better job at writing an article like this. Why? Because a human can probably look at 10, 20, 50 sources.

18:24These LLMs are so powerful. They can look at thousand sources and then make a study out of it if you prompt them to. So it's faster, more efficient. So they can actually outrank us in writing that informational content. But on the other hand, if we're writing something that is only unique to us um like here are five things I learned about operations management managing uh Jolly Search that is uh information only unique to me. And if I give that information out, not only it's going to rank higher, AI will learn that about me that Muskaan is an expert when it comes to operations management. So it's all you know related to different things.

19:10Uh and Greg was saying that uh when he used to be a writer, he used to interview people and this is exactly something I was thinking about that when we do become if and when we again become Jolly Content and we have uh those uh services again, I think we will follow an exact same approach. We're not it's not going to be just a you know a service where the client pays us and we do the articles for them. I think we will have to do interviews with the clients where we can actually get their unique insights. Once we have that data, that'll be more than enough for us to produce pieces.

19:52So, if we've decided, let's say there are 10 articles we're going to produce this month for the client X, we will actually first do a very comprehensive onboarding versus what we do today. we will have to go on a call with the client and do an interview with them uh which will include follow-up questions. So, it's going to become a collaborative piece at the end of the day when it hits the the web. And I think that content is definitely going to outrank any other kind of content because of obvious reasons. And my last um yeah and my last uh best practice for being recommended is having good social content out there.

20:37Uh I think anyone who has been reading a lot about AI visibility would now know this that social and SEO are becoming besties now. Um so social media teams and SEO teams are now I know there are many teams that are working together to produce content that not only goes viral that is not only on brand and will help the company page grow its social numbers but also would help them get recommended in these search results because um if you look at the cited sources LinkedIn, Instagram, Facebook, Reddit all of these sources are cited a lot because um If I say, hey, I'm the best person to talk about AI visibility,

21:21you would think that, hey, Muskaan is a narcissist. But if Greg and Morgan and Gabi, like my entire team goes out doing that PR for me, you would know that, hey, it's maybe genuine. And when you keep saying that for a long time, when anyone talks about AI visibility, my name will automatically come up because I've been associated as an entity to that category. So I think producing social media content that is also optimized for search results is very important if you want to get recommended. So yeah, these are great. These are these are three really important best practices. I you learned these the hard way.

22:01We've been spending a lot of time researching this year. I'm going to put you in the uncomfortable position as we wrap up here to do the crystal ball. This is heading into Q4 of 26. What about early next year? Where where do you see if you want to talk about the landscape or you want to talk about whether the best practices change? It's fully dynamic right now. Things are always changing. What change are you anticipating trying to get ahead of?

22:47I think um SEO's jobs are going to get much more harder and which is a very exciting thing for us at Jolly because uh this is something Greg basically said a few weeks ago before you went on your vacation. You said something, hey, when something becomes hard and we're still able to do it, it means that the value of that service actually increases. So you can charge premium prices for that. Um, so I think the bad news, if you want to look at it as bad news, we look at it as great news for us that hey, our jobs are going to get much more difficult, much more comprehensive because now it's not only Google search that you have to optimize for. There are so many different LLMs that are out there you want to optimize for.

23:24There are social media platforms you want to optimize for. And I think this will keep on increasing and increasing and increasing. um you cannot leave any one platform. Uh one mistake I've seen some clients doing is that they totally um kicked away their SEO and they started focusing on just social and AI visibility that is also not good. Um if you see that we had this news mid August that Reddit citation have dropped uh by a significant amount mid August. I think if there's one learning lesson from that is uh I think it's that you cannot just rely on one platform anymore. If you're only doing Reddit, good for you.

24:08If you want to do Reddit marketing and expand your community, but if you're doing Reddit for AI visibility and search, you cannot put all your eggs in just that one Reddit basket because these things will come and go. So, you have to keep investing in your website. You have to keep investing in digital PR. you have to keep investing at off page and you have to keep investing now in search as well. So I think um this is my take for 2027 that there will no longer be just one platform to optimize for. You have to optimize for so many platforms and it's going to get challenging but it's also going to get equally exciting. Yeah, I love the optimism.

24:52I go figure I agree with myself. Uh, it's going to get harder and it's going to be better for those who want to do the work. So, Muskaan, I think this was full of information. We could keep going another 30 minutes. Uh, we like to keep them pretty tight here uh on the TL;DR, hence the name. So, if people want to follow up with you, you know, I'm not as cool as you. I'm not on all the socials, so maybe you want to list a few. Where can they find you if they want to reach out and ask you some questions? I think LinkedIn would be the best.

25:29Um, if you want to talk AI visibility, if you want to email us or get in touch with uh us through our website, that is also a good platform to get in contact with me. Um, yeah, I wouldn't say Instagram. I'm super fun there. So, just keep that to myself. Don't don't look at AI visibility talks on Instagram. So, well, thank you for coming on and talking AI visibility today, Muskaan. It was awesome and I hope you come back again. Thank you so much, Greg. It was very fun. All right, see you next time, everybody.