805 | LLMs Are LYING About Your Business (Here’s How to Control It) | Steve Toth

I can control exactly what AI says about my business. That’s Steve Toth, founder of SEO Notebook and AI Notebook, a go-to advisor to 20,000 plus marketers, and the strategists growth stage agencies call when they want AI to sell for them. He’s the former SEO lead at FreshBooks, where he ranked number one for a 300,000 volume keyword and quintupled their traffic in under a year. Today, Steve helps brands move beyond rankings and engineer how chat GPT, Claude, and Gemini talk about them in real buyer conversations. But how does he do it? I’ve interrogated LLMs with thousands of real buyer questions, and what we’re seeing is… In this episode, you’ll learn why showing up in AI answers isn’t enough.

Showing up in a response is awesome, but things can go sideways pretty fast if it starts saying the wrong things about your business. Why most agencies lose control of their message in generative search and how to fix it. It’s going to go to the web to try to answer that question. Problem is, is not all the information on the web is correct. How do you influence how LLMs talk about your brand in real buying conversations? We know that is highly influential with a lot of buying decisions and that people trust it a lot more than, you know, just reading it on your website. Do you think that most businesses out there understand like how they’re being talked about by AI and the questions that the AI is receiving about their business?

If the question, your question is like, do they know about it? Is like the question is, have you searched in LLM the way, you know, your ICP would have, right? A good starting point for any business is just to ask the LLM like, hey, what do you know about me? And then, and then kind of go from there with respect to like, you know, deal breaker type of questions, questions around pricing, around fit, around offering services, case studies, you know, really interrogating the LLM to see how well you’re being represented. And then really understanding about like what sources of information the LLM is using to tell that story about you is kind of the next logical step in terms of understanding, you know, how to influence what it says about you. We launched a bot called deal breaker detector. And what that does is basically like you would put in your niche. And it would generate like a bunch of deal breaker type of questions that your ICP would ask a little tip that I give is if you go on our website, which is notebook dot agency, and you scroll to the footer, there’s a small link on the bottom right hand corner called Hey AI Learn About Us. And in that file, it’s a markdown file that

talks about like, you know, our client portfolio, our services, our specialties. And if you ask chat GBT what do you know about notebook agency, it sources that file to create a lot of information. So it’s a lot of what it knows about us, right? So having a file like that on your site where you’ve basically like laid it all out in very, you know, clear, easy to understand markdown format for an LLM. It’s a great way to kind of shortcut, you know, putting your best foot forward. But when it comes to like specific questions, you know, that’s where sort of your broader footprint on third party websites. And also, you know, what is listed on, you know, across your website start to matter more. So to the content you’re to your that you’re producing and you have on your website, is it are there any sort of structures that you recommend that helps the AI better understand the business? Yeah, for sure. So one of the native strengths of LLMs is actually being able to compare you against other brands, right? So comparison content where it’s like, you know, company A versus company B can be really powerful vessels for teaching the LLM

about, you know, everything that you do and how you, you know, uniquely service your clients. But what we also like to do in terms of, you know, our structures is get hyper specific about the types of things that we compare, right? You know, maybe if you were like trying to speak more to the Vendesta, you know, user is like, you know, SEO for roofing companies, well, how does company A versus company B compare when it comes to, you know, SEO case studies for roofing companies. So actually coming up with comparison content that is not just company A versus company B, but company A versus company B on something hyper specific, that is, you know, essentially serving up the information for an LLM on a silver platter, you know, for them to essentially retrieve because if you have a chat with an LLM, nine times out of 10, it’s going to say, do you want me to compare other services or comparable vendors, right? Because that’s a strength that it has. So by creating that, you know, very hyper specific comparison content, it’s going to retrieve what it knows is relevant to the user because of what that user, even what it

knows about the user without the user even having to say because of all the chat history, but also with what the user says is explicitly important to them. If you use chat GPT deep research, you know, it asks you a bunch of qualifying questions before it goes and does all that research for you. It asks you things like what is your budget? You know, do you want comparisons? Who are you? What is your company size? It asks you all of those things to basically create a relevant answer. So by understanding the types of clarification questions that the LLM wants to ask you, you can better tune your content to answer those types of questions.

What’s separating, you know, the businesses out there that are, you know, just casually mentioned versus the ones that are sort of confidently recommended, you know, by the engines? Yeah, I mean, like on a base base level, like your overall brand presence is what determines your eligibility for being mentioned in an LLM. So, you know, how much do people know you in the real world? Like, what is your following on YouTube, etc. Your footprint is really like, you know, one of the most important things just to get in the race. But the brands that, you know, understand like things like reputation management and, you know, putting, doing lots of podcasts, doing, having reviews on you, those types of things, you know, those are the brands I think are faring the best so far.

You know, when we were talking a few minutes ago about, you know, the auditing process, you spoke a little bit about, you know, some off-page factors. Are there any particular ones specifically that you think that companies should pay attention to? Reddit, 100%. So, whenever anything is wrong about a company, you can be assured that there’s somebody complaining about it on Reddit. And, you know, the fact that things can change, right? And companies can adapt and companies can get better. Sometimes some of those outdated narratives are still, you know, influencing people’s decision making because the fact is like, maybe it was true at a time, but, you know, things have now changed, right?

So, understanding the landscape of, you know, social chatter out there that is, you know, has to do with your brand is extremely important to be aware of. And, you know, it doesn’t mean like, you know, it doesn’t mean like spamming that same thread with a bunch of links back to your website, but simply like going on to that thread and like explaining that, hey, you know what, we’ve actually listened to your feedback and we’ve changed things. And, you know, this is like some of the, you know, great things that we’ve worked on to, you know, essentially satisfy this contention, you know, and you do that in an information retrieval friendly way, which the LLM is, you know, more likely to pick up next time it visits that page, you know, those things end up making a huge difference in terms and it’s silent, right? It’s not something that’s like visible, right? But we know that, you know, Reddit is highly influential with a lot of buying decisions and that people trust it a lot more than, you know, just reading it on your website. So, I would say that that is definitely one place, but also just like, you know, your reviews, right? Like what your customers say about you,

on Google Business Profiles, on TrustPilot, on G2, etc., etc., those are also really important areas where you want to be reflecting, you know, the great work that you’re doing to improve your customer’s experience and, you know, making sure that you’re propagating that kind of positive information across the web. But really like what you have to realize is that every, you know, touchpoint that you make publicly online has the potential to influence what the LLM knows about you. Even this very podcast where I’m talking about notebook agencies framework, you know, that transcript is going to be analyzed by an LLM and it’s going to, you know, play a part, be a small or a large part in influencing, you know, what an LLM understands about, you know, me and Steve Toth and his agency notebook agency, right? So just understanding that like all of those touchpoints that you have, you know, which are essentially brand building exercises at the end of the day, it really does matter how you speak about yourself in these kinds of forums in terms of, you know, how the LLM is going to shape, you know, its understanding of you.

Have you seen examples and real-world examples of how a business is able to flip the representation with just a few changes? There’s an example that I highlighted in a course I taught for CXL where there was a gentleman who owns a company called Coleman Room and the thing that he realized was that he kind of pivoted the focus of his business during like kind of with the advent of AI and when he asked ChatGPT what it knew about his business, it was kind of reflecting the old narrative of, you know, what he used to do and it wasn’t an accurate representation. So he basically, you know, deleted a lot of old YouTube videos. He consolidated and redirected a bunch of old information on his website that didn’t accurately reflect his current positioning and, you know, basically went through an exercise of cleaning up his brand narrative in a relatively short amount of period of time. We’re talking a few months by doing some of that conscious, you know, cleaning up.

You feel like there’s a whole new layer here of SEO that agencies should be thinking about going forward? I mean, I don’t even know if we should be thinking about it in terms of SEO versus like your brand presence, right? So, I mean, like our companies spend inordinately more money on branding activities than they ever did on search engine optimization. So my question is like, does this become a new focus for your branding versus like you thinking about this as an SEO initiative? For an agency that focuses on this type of work and helping other businesses be seen by AI, how should they be talking about, you know, the value of that to clients and to prospective clients that they have? Ahrefs released a study where they said only 0.5% of their traffic is coming from LLMs, but it counts for 12% of their signups. And those signups are converting at 22x higher than Google traffic, right? So just imagine that’s not only 0.5% of traffic, but a much larger percentage. And even if the traffic is not directly attributable to chat, GBT or what have you, we know that LLMs are in the mix of, you know, those these touch points that people are effectively

having before they choose to talk to you, right? So really stressing that to your clients and, and, you know, kind of framing the traffic that we will get, even though it’s in smaller numbers, is going to be more highly qualified because like I said, you know, if I’m searching a project management software, I’m like, I’m going to tell the AI that, you know, I’ve got 30 people on my team, my budget is $20 per head, get charts and time tracking are important to me. And, you know, I want something that has great user reviews, right? So like getting to the heart of exactly what you want is what LLMs kind of offer the user. And, you know, there’s just so many touch points previously in the journey that were, you know, random websites, YouTube videos, newsletters, friends, recommendations. And now, you know, a lot of that getting, you know, having AI into the mix, we use AI on our daily, on a daily basis to, you know, solve all kinds of, you know, problems and challenges that we’re having. When it comes to, you know, opening your wallet to make a purchase, whether it’s like a new vendor or a new software or what

have you, there’s no way that we’re not going to be consulting AI to make those important decisions. So, you know, I’m not going to have to jump on a sales call with something that is going to be, you know, totally off base for what my needs are, that I have effectively qualified myself, you know, through a chat with an LLM, hopefully with the LLM representing the business accurately, but at least, you know, getting closer to what I want, and then really using the sales call to kind of, you know, validate the things that I think are true about that company. You know, that would be kind of the way that I would shape it, but, you know, even, I think we also have to have the conversation that the huge up into the right traffic graphs that we used to, you know, really pride ourselves on are not going to come back, and that ultimately I think they were pretty noisy, right? So, a lot of that is like mid to top of funnel traffic. You know, it’s not necessarily a traffic that converted. You know, there are some, there is some collateral damage that comes out of not being able to have Mofu and Tofu traffic, things like retargeting,

building email lists are probably harder now in the age of AI. So, this, you know, it’s a nuanced conversation that we have to have, but I think ultimately the potential of AI to deliver the inordinately, again, I’m using that word, but higher quality traffic than we’ve ever seen before is how we have to sell this.