Episode Transcript
Kyle James (00:00.876)
Hey, welcome to the AI Chronicles podcast. I'm your host, Kyle James. And today we'll be talking about how an AI powered organic search company called TACO is using AI inside of their own business. And we'll share the exact steps that you can take in order to implement AI for yourself. Now, before I dive into that, listen closely. Are you looking to implement AI inside of your own company or maybe just struggling to get your AI to stop hallucinating? Speak to GPT trainer.
GPT Trainer literally builds out and manages your AI for you, eliminating hallucinations for good. Go to gpt-trainer.com. I promise you, it'll be the biggest time-saving decision that you've made all year. Trying to set up AI on your own is like trying to build a house from scratch. Sure, you could do it, but the time of frustration is gonna take you to get it finished. It may not be worth it. It's a thousand times faster and safer to hire professionals.
schedule a consultation today. Once again, that's gpt-trainer.com. Today I have with me on the show Chirag Kulkarni, who is the CEO and founder of TACO, an AI assisted organic search partner that helps businesses rank and drive revenue through Google, chat GPT, and perplexity and many others. Chirag has helped companies from startups like Evi to Fortune 500 companies like Expedia and LinkedIn.
drive more revenue through organic search. It's gonna be a very interesting conversation. Hey, Trog, welcome in. How are you?
Chirag Kulkarni (01:35.108)
I'm good Kyle, thanks for having me.
Kyle James (01:36.782)
Yeah, man, so excited about this. Now, you've been, y'all have been busy. Like, it's been a growth season for your company. Like, tell us a little bit of backstory here. How did taco come to be? And maybe a little bit about yourself too. So we have, audience has a little bit of context.
Chirag Kulkarni (01:51.682)
Yeah. So my, my quick background is, I got my start in the SEO world about 10 years ago. I worked with a guy named Neil Patel for a little bit. And that's, that's really where I learned SEO kind of firsthand learning from somebody who had kind of skilled his own sites, two millions of visitors a month.
He ended up introducing me to a number of companies. And through that process, I ended up becoming the CMO of this business called Medley Pharmacy. And it was really there that I got to kind of see firsthand the power of channels like SEO within a business as a chief marketing officer, right? So.
I was managing a budget. was working with our sales team and different stakeholders to basically help get more business. And one of the things I historically found was that most of these agencies that I would hire, they tended to be extremely focused on traffic and not actually what matters, which is conversions and how do those conversions result in revenue.
And the second insight that I had was that AI that is more or less more common today than it was, say, three, four years ago, was going to completely change the way that we work, but also the way that we search.
And so I felt like it makes sense for me to leave this business and go build a business that I wish I could have hired as a chief marketing officer. And that's why we started TACO. you know, like you rightly said, our job is we use AI assisted workflows, agents with human in the loop.
Chirag Kulkarni (03:23.748)
to drive kind of outsized returns from an AI search perspective. So what does that mean? A business comes to us and instead of just giving them tooling, we ultimately help them drive outcomes, right? So when it comes to content production, it comes to determining, hey, which citations do we need to optimize for? How do we reverse engineer those to drive search performance? So the lens becomes, hey, how do we help traditional businesses?
navigate from traditional SEO to AI search, right? So that they can optimize for chat GPT for perplexity. And also for newer companies, how can you leverage these AI search strategies to get ahead of their competition as maybe some of the more legacy businesses struggle to make that transition. So that's what we do on a day to day basis.
Kyle James (04:06.957)
Yeah, so that's been such a huge conversation topic in the SEO world from what I've been hearing. It's just because the AI has been playing such a big role in the way people search for things. Getting enlisted top page on Google versus getting enlisted in a large language model. There's kind of that tug of war that's happening. Where do you see it kind of going, I guess, in this case? But also, is it a both and or? Okay, be in this more or be in both or should we start transitioning? Because there's some companies who spend
millions of dollars right on getting on that top page, but now it's like where do they spend that funds next? That's gonna get them that good return.
Chirag Kulkarni (04:45.944)
Yeah, I think the way that we think about this is that, you know, and Forrester and other groups have said that in the next three to four years, traditional search, right, the way that you and I know it, you go to Google 10 blue links is naturally going to start to decay and it already has started to decay. But what's going to be replaced by to your point is it's going to be, you're going to go into an LLM and you're going to ask the LLM a question, right? What are the best running shoes?
in the US that I can purchase. Nike shows up or Deezus shows up, right? Maybe Puma does not show up, right? So the belief here is that we need to continually optimize for traditional search while giving companies the leg up in the strategies and tactics that work today, as well as in the future and as in kind of in a test and learn capacity. So for example,
You know, we have a call it 20 plus businesses that we are working with and we are constantly learning what are new strategies that are working well in healthcare that are working differently and say AI software applications that are working differently in FinTech. And so we can apply those learnings and tests and learn to help companies get a better outcome from an AI search perspective. So that's the way that we think about it.
Kyle James (05:59.438)
Yeah. And like, so I'm not sure how long like how long AI has been integrated. Like you're obviously you're using AI over at Taco and I love what you guys are doing and the backstory here and like specifically though, like, I mean, I see where it's going now. Like initially was it like a, was it because you saw the change that you made the change to the AI or is it like a, Hey, look, I'm seeing, you know, let's just, Hey, AI is new. Let's go ahead and start exploring it. And then you discovered, Hey, here's some of those challenges that are associated with it.
And maybe even future challenges more. kind of like paint that picture, like what was going on when you first made that transition on the AI side.
Chirag Kulkarni (06:35.438)
So the transition was quite natural, Kyle, because what ended up happening was we ended up hiring individuals like specialists that were doing specific tasks and work. And what I quickly found was that, we can kind of build AI workflow, let's build agents, even as simple as like, know, GPTs, custom GPTs, right, that we can use to just make our lives a little bit more efficient.
Instead of, writing meta descriptions, title tags, H1s from scratch, can we determine what the primary keyword is added to a custom GBT? It spits out. Was the V1 of where we started from an AI search perspective? In terms of how do we use AI internally, to your earlier question.
Where that evolution has kind of taken us. And I think the way that we think about this is what are all of the most tedious tasks that in the business we have to do and where can we build AI assistance to allow us to get that much more throughput so that we can get more better, faster outcomes for the companies that we partner with. So for example, the content writing process, everyone knows you can use a Jasper or a copy AI. You can put in a particular keyword. It spits out a piece of content.
But what we originally thought about is like, the content writing process, there's a process to writing content. And what AI can do is, you know, be very intelligent and unique about figuring out, for this particular keyword.
What is the intent around it? Is it commercial intent? Is it informational? What do I actually need to include in each of the paragraphs? Right? What is the H ones? What are the H twos? The H threes, H fours, which is the structure of the actual content.
Chirag Kulkarni (08:17.529)
This is the kind of stuff that we use AI for. So it's not like, hey, we're just going to generate content using AI. It's that instead of having people write every single word, can we now dictate what content needs to be in each of the sections to then get passed on to an editor who can craft the piece, who can add their unique insights in that specific topic?
to again create super competitive content. Instead of them say doing three pieces of content a day, they can do six pieces potentially. I'm making up the numbers here, but that's the idea. It's the same thing from which actual queries do we need to go after for collections pages or product pages in e-commerce? So how do we reverse engineer those into questions that are being surfaced on LLMs?
Kyle James (08:41.357)
you
Chirag Kulkarni (09:02.275)
These are just a couple of examples of how we're leveraging AI assistance to make our work and throughput faster for the individuals who are actually executing the work.
Kyle James (09:12.063)
Yeah, that's exactly some of the things too, I've been seeing, even the conversations I've been having is people create different types of AI agents and it's okay, who's using it? It's just like a lesson planning agent for education. like, well, who's using this lesson planning? It's like...
Well, it's those who are creating the lesson plans. who would have thought? Like, who's using the marketing AI agent? Well, it's the marketing team. So it's it's amplifying to like what you said though, Travis, like it's the taking that the initial output and then going, okay, how do I get it to have unique insights by putting it in the hands of really the human expert? That way it really goes that to that super competitive nature versus just, hey, if you're just using AI basic output, like you're not, that's not gonna be.
the unique, that's not going to make it tailored, but it can certainly help the team escalate and kind of shifting gears here a little bit, because I know, you know, I want to go back a little bit on the, on the AI and being listed in these LLMs, like that's such a big piece. think a lot of people are curious about it. Like walk me through kind of, you know, the AI step-by-step for like a client you're working with, like how does that process look like where they go from being listed on Google, right? Or the top page of whatever the big search engine is.
to being listed in perplexity and Grok and OpenAI and Claude and Throb, like how do you get to that point?
Chirag Kulkarni (10:29.465)
Yeah, it's probably a longer answer and is kind of dependent on the actual query. And the reason I say query is because you are searching for questions in L-Lebs. In Google, you're typically searching for keywords. But the way to kind of think about this is that you want to first start determining, when I search for a particular question, I have to figure out where is that question going to come from? For example, let's take the Nike running shoes example. So men's running shoes is the keyword.
Somebody is not going to necessarily go into a chat GBT and search men's running shoes. They may search what are the best running shoes for someone who's about to run X race, right?
So what we have to think about is like, are those actual questions that are people are searching for based on the keywords that are actually driving us revenue? It's one thing to be listed as, Hey, like the LLM is servicing me for what is taco, right? Or who is Chirag. But that's not really what we want to do. We want to think about what are people searching for that they're interested in buying a product or service for, and how do I reverse engineer those into questions? So that's kind of step number one.
Step number two is that once we actually input those into LLMs, as you know, the answer that we put in the logged in version, the logged out version are all different. So we have to run several tests to determine what is the actual answers that are coming back.
And then the third thing that we need to do is we need to determine, what kind of data are LLMs referencing? What citations are they pulling in from? Is it Reddit? Is it Wikipedia? Is it certain owned websites? And that helps us determining, OK, what's the strategy that we need to do in order to drive optimizations from an LLM perspective? And then we kind of build out a custom plan for a set of keywords on a monthly basis and sort of questions based off of those. And that's how we kind of attack the LLM search.
Chirag Kulkarni (12:25.273)
game. So that's kind of at a very high level how we think about it.
Kyle James (12:27.861)
Yeah. Yeah. So like on the, like you mentioned, like, cause I noticed too, like when I'm on different, you know, LLMs, I'm searching content and like one of the things, depending on which model I choose, right. It'll say like, it's referencing these three different types. Like I see Reddit on there. I see LinkedIn. I see maybe like Google or whatever. So in this case, like it's, is that like kind of tell a lot of companies to say, okay, you need to be targeting like blogs or your data on those websites because that's naturally what the LLM is pulling in. Is that typically?
Chirag Kulkarni (12:51.043)
Yes.
Chirag Kulkarni (12:56.835)
Yeah. So think of it like this. The, the question, what are the best payroll software companies and the question, what are the best running shoes for men who are about to run X race? They're going to be pulling from different data sets and different sources. Best payroll software or maybe a bunch of articles listening to top 10 payroll software companies. This other one may be Reddit, right? Because they really want unique insights around what people are saying on Reddit on Quora.
Kyle James (13:22.432)
Hmm
Chirag Kulkarni (13:25.005)
So you have to understand what are those patterns and what are those patterns specific to the questions that are being searched for on LLMs.
And then you need to be found on those particular places. So for example, you have to still go and optimize the content, which is very traditional to SEO, but you need to do it in a way that the structure is similar to what the LLM is spitting out. You need to make sure you're adding in relevant data, right? Because LLMs love to cite data. And then you need to think about, okay, if TikTok is being referenced in this piece of content, then I need to go create a TikTok piece of content around this topic in order for me to then get cited in the LLMs.
Kyle James (14:01.579)
Wow, that's, I think, I think, don't know, my mind's kind of blown right now within this. It's like, it's like, it's, what you're telling is like, okay, I, why didn't I, why didn't I not see that? I think a lot of companies feel that way. It's like, like going in almost kind of reverse engineering, okay, what are they referencing? Because I think the game of just, Hey, putting blog posts or Hey, Facebook marketing or whatever it is that they're doing. No, it needs to be more original content. And if original content is only in, you know, two or three areas,
Hey, it might be time to create a TikTok account, maybe an Instagram or, I don't know, Snapchat. I don't know if it Snapchat, but like it's expanding, you know what mean?
Chirag Kulkarni (14:36.003)
Yeah.
Yeah, exactly. SEO used to be primarily your site, your owned assets on your website. SEO has moved to now being much more about your digital footprint. And Kyle, what I think the transition that's happening that people are not ready for is cost per clicks are increasing every single year, which is basically the money that you're spending to get a click from Google, from Facebook in terms of impressions for Facebook.
So that's increasing, right? Well, what I think is going to be fascinating and where this serious opportunity lies today is because to get mentioned organically in chat GPT.
is a real opportunity right now, right? Because if you can capitalize on it, then you're riding this wave as we've gone from, say, 100 million searches a day to a billion searches a day on ChatGBT and call it five months. Google for context has about 14 billion searches a day. So we still have a little bit of ways to go, but ChatGBT is growing really, really quickly. People use different products in order to find their solutions. And so I think the opportunity here is that
Kyle James (15:22.541)
Mmm.
Chirag Kulkarni (15:46.935)
organic presence on LLMs is going to be massive from an arbitrage perspective, as well as the businesses that take advantage of it today are going to get an outsized opportunity and they're going to be able to efficiently acquire the right customers leveraging these channels versus waiting longer and longer. And it's kind of like when Google just launched, right? People were really taking advantage of cost per clicks then. So that's the way that, know, businesses can also think about it.
Kyle James (16:13.581)
So I'm gonna backspace, I'm sorry I'm all over the place here. I keep going back, so like you said earlier, and it's just got me curious, like the question part, like you mentioned like the example of searching for like the best Nike shoe or Nike shoe for summertime, I the search. That's the question, say people are asking, getting listed in LLM. So essentially you're taking that question and building out content that answers the question? Is that like how, that's the angle you would attack it.
Chirag Kulkarni (16:40.364)
Exactly.
Kyle James (16:43.617)
Is that?
Chirag Kulkarni (16:43.639)
More or less. Yeah. And, know, like, again, in the example of like best running shoes, right. Versus best running shoes to run the New York marathon. There are going to be relations to those two queries. It's not that they're completely different answers. Maybe what we do in that case, and I'm spit balling here, but like, maybe what we do is we add a section on our content around the New York marathon.
Kyle James (16:54.893)
Mmm.
Chirag Kulkarni (17:05.849)
Maybe we find that, in Reddit, there's a specific subreddit all around the New York Marathon and people are asking for what are the best shoes? Have people tried this version of ASICs versus this version of Puma versus this version of Hoka? And we go in there and talk about our new brand in those Reddit threads. So now LLMs are starting to notice like, okay, for this new business, this website talks about...
Kyle James (17:07.821)
Hmm.
Chirag Kulkarni (17:29.305)
the New York marathon and running shoes. This Reddit thread, they talk about the same thing. This listicle, they talk about the same thing. So like these are all related and that's what drives LLM relevance. And that's what our job as a business is to do is to not just report on, hey, Kyle, you show up in LLMs this month of times. Our job is to do that. And our job is to actually determine how do we actually improve this? And can we actually go and execute on doing the activities that I've just mentioned here?
Kyle James (17:55.886)
Yeah, that's cool. So like it's a picture it now or it's like the running shoes, right? The first made the first part of the sentence is like, you know seven words, but then that last four words you can almost take seven of the most common types of questions after that first part. Does that make sense? I'm almost like trying to do like if I had a whiteboard right now, I'll be writing it down. But like that's it right the first half is is the staple but then the last variable of it you could put those four different variables into say one.
piece or multiple pieces of content. Because now it's going to be, in a sense, theoretically would get you towards the top of the LLM.
Chirag Kulkarni (18:33.573)
you
Exactly. And as you do these activities over time, you'll notice LLM visibility is increasing. And so what's ended up happening is more and more businesses are being offensive about content, right? Because the more content you create, the more opportunities you have for LLMs to ingest data and be able to ultimately understand your business versus your competitors versus the approach that, you know, about two years ago, some people said, which is SEO is dying. It's best to pull back on budgets and SEO and reallocate them
towards paid, right? So there's, you know, different schools of thought, but like the businesses that I find are taking advantage of this are doing the offensive approach for the reasons mentioned. I'm also biased though, let me be clear, but outside looking in, if I'm a CMO, this is the way I would think about it.
Kyle James (19:05.442)
Mmm.
Kyle James (19:14.573)
Yeah, of course.
Yeah, yeah, no, it's cool. It's like, it's wild to think like how much more important it is like to be generating content for every single business out there because you're just you're really kind of dumping it into this massive like fuel engine of an LLM and really you're hoping, hey, you're hoping your your sponsor logo gets placed on the front of the windshield for everyone to see like as it begins to move forward. And that's the visual here I'm seeing. and so tell me a little bit here. Now, obviously, you guys are doing a lot of taco.
Chirag Kulkarni (19:28.09)
Yes.
Kyle James (19:45.299)
A lot of plans ahead, things will change over the next couple of years, months, right? But where do you see AI playing maybe some of biggest role in your operations in the near future?
Chirag Kulkarni (19:58.167)
So we have a of an individual group called the Growth Engineer. And their job is to literally think about all the areas in the client business type relationship and execution from an LLM search perspective and figure out what are the AI workflows, toolings, tactics that we need to build in order to make our lives more efficient, but also more effective for the customer.
And so what we are actually doing Kyle today is we are thinking through every single step of the process to allow us to rank competitively on AI search, testing, implementing, reporting, and building those relevant toolings and kind of workflows in order to get there. So the way that I see it is, that
because of how dynamic marketing is to your point, the LLM or AI is not going to continuously learn what strategies are working out versus not, right? It's our job to test and learn those things and use AI to kind of drive more throughput. So I kind of see this in like a test and learn capacity and to continually build these kinds of workflows and tactics to get there. So.
Kyle James (21:11.625)
Yeah, that's cool. I like the process that you have. Like it's not just like a, well, I'll figure it out. How does it go? It's no, no. Here's the strategy behind it of how we make sure that like we're not missing out when because it is coming fast. It just like I said, it's a wave. If you don't catch that next wave, the next set might not be for another five minutes. But we'll be prepping. That's a surf. That's a surfer cue. I'm a surfer guy. So had to. You said wave and I was like.
Chirag Kulkarni (21:30.499)
Yeah.
Kyle James (21:35.405)
I gotta talk about ways, man. So, okay, and as we start wrapping, Triog, this has been a solid conversation. I know this has probably been very enlightening for lot of business owners out there who are listening in. Where can people go to learn maybe little bit more about you and then maybe a little bit more about taco that you recommend them checking out?
Chirag Kulkarni (21:36.576)
Hahaha
Chirag Kulkarni (21:52.345)
Yeah, I think the two great places is like we publish and give away a lot of stuff. Like for example, this week alone, you know, we gave away two or three kind of interesting things from like, Hey, here's like a GPT that we've given me access to around LLMs.txt, which is like a new way for LLMs to actually read your content more effectively. And it's considered more effective than robots.txt. It's a test.
It's not like fully endorsed by Google or LMS, but like that's one example that's on LinkedIn. And then, you know, if, this is interesting and you kind of want to think about how can you use AI search to kind of like drive more throughput or optimize for it accordingly. And you're looking for beyond reporting. We partner with a lot of the tools out there to deliver this like scrunch, for example, but tacos of TACO like Taco Bell or, or tacos in Austin.co.co. So you can go there.
and book some time with the team and we'd love to chat with you.
Kyle James (22:49.515)
Yeah, absolutely. And appreciate you being on the show, Chirag. It's been a pleasure having you on, brother. Hopefully we'll see you on the next one. Cool. And thanks, everybody, for listening in. Remember, please, if you're looking to implement AI into your business today, please don't try and do it yourself. The time and stress that the AI could cause, it may not be worth it. So please schedule a call with GPT Trainer and let them build out and manage your AI for you. Once again, that's gpt-trainer.com.
Chirag Kulkarni (22:58.073)
Sounds good, thanks Kyle.
Kyle James (23:17.346)
Have a great rest of your day and looking forward to seeing everyone on the next episode of AI Chronicles.
