Podcast: Play in new window | Download
The way people research businesses is changing fast – and AI is increasingly shaping those decisions before someone ever visits your website. Tom Rudnai shares research into how large language models influence buyers throughout the customer journey, and why traditional SEO thinking may no longer be enough when it comes to visibility, positioning, and staying relevant as AI becomes part of the buying process.
Why Your AI Visibility Strategy Is Only Seeing the Tip of the Iceberg
If your current plan for getting your business “seen” by AI systems is trying to rack up citations in ChatGPT responses, I want to push back on that a little. Not because citations don’t matter, but because my conversation with Tom Rudnai of Demand-Genius made clear that citations are capturing maybe 16% of AI’s actual influence on your buyers. The other 84%? It’s happening in conversations your analytics will never see.
The Research Behind the Claim
Tom’s team at Demand-Genius didn’t just theorize about how AI systems behave across the buyer journey. They ran a large volume of structured prompts spanning awareness, consideration, and conversion stages, then analyzed the responses in detail. The goal was to test whether the exploding category of AEO (Answer Engine Optimization, sometimes called GEO, Generative Engine Optimization) advice was actually built for how AI works, or just SEO advice wearing a new hat.
AI searches the web 0% of the time during awareness and consideration stages. It’s working entirely from training data. It’s not going to Google. It’s not looking for your blog post. It is drawing on what it already knows, and using that to help your potential buyer frame the problem, build requirements, and start narrowing their options. Only at the conversion stage, when the query intent becomes decisional, does it start searching, and even then, only about 48% of the time.
So if your strategy is optimized for citation and traffic, you are only showing up, in the best case scenario, in roughly half of the bottom-of-funnel queries. The full funnel influence is largely invisible.
Search Is a Directory. AI Is a Thought Partner.
One of the clearest frames Tom offered is this distinction. Search takes your keyword and treats it as an isolated event. It finds the best matching document. AI is doing something much more complex. It’s building a bespoke answer for that specific person’s specific situation, incorporating everything they’ve told it across the conversation.
A single keyword like “best CRM solution” might map to a handful of content pieces you could create. Tom’s research suggests that when you account for the range of job titles, criteria, use cases, and contexts people bring to that same query as an AI prompt, you’re looking at roughly 22,500 prompt variations. You cannot one-to-one-match content to prompts the way you could with keywords.
That’s not a reason to panic. It’s a reason to think differently.
Where AI Influence Really Lives
Tom used the analogy of the old sales world problem of the RFP. There’s a well-known cliché in sales: if you just receive an RFP cold, you’ve probably already lost. Chances are one of your competitors has been talking to that buyer for a year, helping shape what they think good looks like. They influenced the questions. They wrote the rubric that leans into their strengths.
AI is doing the same thing at scale, at the top of the funnel, for every buyer in your category.
If you’re only optimizing for visibility at the conversion stage, you’re filling in an RFP your competitors shaped. The real opportunity is getting upstream and influencing how problems are framed before buyers ever ask for recommendations.
What to Do About It
Tom’s framework centers on three levers: your positioning, your content, and your reputation (what others say about you on third-party platforms like LinkedIn, Reddit, and G2). He’s clear that we don’t yet have precise data on exactly how much weight each carries, but the goal is to make all three consistent. The more alignment there is between what your positioning says you are, what your content demonstrates you know, and what your reputation reflects, the less room there is for ambiguity. And ambiguity is where AI either makes things up or lets competitors fill in the blanks.
On the content side, the most useful concept Tom introduced is information gain, which he breaks into three levels: interpretive gain (a new angle on an existing idea), empirical gain (original data and research), and conceptual gain (a new framework that reshapes how a category thinks). The argument is that AI can already summarize existing knowledge perfectly well without you. If your content is just doing the same thing, it doesn’t need you. Content that creates information gain is what earns inclusion.
For a large enterprise, this might mean funding original research. For a smaller business, Tom was honest that the scale of your effort should match the complexity of your purchase decision. If you’re selling something with a short, straightforward buyer journey, the existing visibility-and-citation advice might serve you fine. The more complex the category and the longer the sales cycle, the more important it becomes to do the harder work of shaping how your market thinks.
The First Steps Worth Taking
Tom’s recommendation for where to start comes down to a simple diagnostic before any strategy work. First, figure out whether AI is even a meaningful channel for your buyers. Not every business needs to optimize for ChatGPT. If your ideal customers are not heavy AI users in their purchase research, chasing AI visibility is a branding exercise, not a revenue driver.
If you’ve confirmed that yes, your buyers are using AI in their research process, the next step is understanding how they’re using it. What are they asking at each stage? That context shapes everything else. And then, only once you know that, start matching your content and positioning strategy to those specific stages.
The final thing I’d add from this conversation: stop treating AI visibility as an SEO sub-problem. It has overlapping fundamentals, especially around authority and expertise, but it requires its own strategy, its own metrics, and its own content approach. The businesses that recognize that now are the ones shaping the category conversations that will influence buyers long before any recommendation gets made.
Next Steps
- Audit whether your current AI visibility strategy is based on citations and traffic metrics alone. If so, you’re capturing a small fraction of AI’s actual influence.
- Survey or talk to a sample of your best customers. Ask them how they use AI in their research and decision-making process.
- Look at your existing content and ask: does this deliver information that AI can’t generate on its own, or is it summarizing knowledge that already exists?
- Pick one content piece to develop at the interpretive or empirical information gain level, something that brings a new angle, new data, or a new framework to your category.
- Align your messaging across your website, your thought leadership content, and your third-party reviews and profiles to reduce ambiguity about who you are and who you serve.
Transcript from Tom Rudnai’s Episode
Rich: My next guest has a decades experience in AI and startups across media, SaaS, and consumer technology. As founder of Demand Genius, he spent the last two years researching how LLMs actually behave in complex buyer journeys and building the tools to help marketing teams act on it.
Today, we’re going to dig into those findings and how they will impact your online visibility, with Tom Rudnai. Tom, welcome to the podcast.
Tom: Great. Thank you for having me. Pleased to be here
Rich: What led you to start exploring how AI is changing the way people discover and evaluate businesses?
Tom: Oh, too much free time on my hands, mainly. I mean, we came into this a roundabout way, so I feel like the best startup founder story would be if I say that it’s been my calling since I was 10. But since it was only invented about three years ago, I don’t really have that available to me.
I mean, we’ve been working in the content space for a while, so my background and my co-founder’s background was in the digital media space helping big media companies figure out how to sell consumers subscriptions at a much higher volume and velocity.
And then that’s what ultimately led us to found Demand Genius, where we were looking at B2B content strategies, and that just happened to coincide with the kind of rise of generative AI and this whole kind of massive, what has literally been called a gold rush of people trying to figure out how to get AI systems not just to cite them, but to mention their brand and to reflect their brand in the way that they want to. And I think there’s two slightly different goals there, which I’m sure we’ll get into today.
Rich: Excellent. Yes. And that is what I wanted to get to. But first, tell me about the research you did. How did you gather the data?
Tom: So one of the things we switched onto a little while ago, or one of my firm beliefs in the world of marketing is that AI should be a force for better, more quality, more diligent content rather than just more content. So we were kind of thinking, okay, how can we use AI to do big pieces of research or something really cool that without it we couldn’t have possibly done?
And we realized the easiest way is to run a load of prompts. And we have some pretty good technology here at Demand Genius to help analyze the sentiment of those outputs, to analyze content itself, and things like that. And we kind of realized we could combine those things, so we run an awful lot of prompts.
But specifically, we wanted to understand how AI responses vary at every step of the buyer journey. I guess the hypothesis that I had was that all of the AEO AI search optimization you see – AEO and GEO are the two acronyms that you’ll probably hear me use interchangeably over the course of this conversation. All of the advice sounded suspiciously like SEO advice, that this brand-new problem was just being jammed into, which didn’t seem quite right.
It’s a fundamentally different technology, so as convenient as it would be if the way that you optimize for AI search is exactly the same as you did traditional search, that seemed pretty unlikely. And it seemed that it was all focused on visibility at the point of conversion. So it’s focused on when someone puts in a prompt of which CRM should I buy, how do we get that to say Salesforce, or how do we get that to say HubSpot?
But the difference between search and AI is search operates as a directory. So it takes that keyword, that search intent, and it treats that as an isolated event. AI doesn’t. That final question of which CRM should I buy, there’s a long thread of conversational context, of problem discovery, requirement building, that leads into that. And we felt that there was something really missing in AEO strategy because that was being ignored.
So that was where we set out to literally, in a very structured way, run prompts across awareness, consideration, conversion, and understand how responses vary at each of those stages. And we saw that it’s pretty revealing as to how over the course of that process with one of your potential buyers or one of your listeners’ potential buyers, the AI responses are going to evolve, and it’s going to help them converge on the right ultimate.
Rich: So when we talk about things like product buy, AEO, GEO or whatever acronym ends up winning out, what are we really describing in plain English?
Tom: We are describing the process of influencing the way that AI systems present your brand. When they present them, who they present them to, and what they say about you. I think it’s an unhelpful distinction that some people do distinguish between AEO and GEO. GEO being the more deeper conversations that you might have with an LLM.
Whereas AEO being like Google’s AI overview, literally quick answers to quick questions. That seems to be the distinction that has emerged. Generally speaking, I think you can treat them as interchangeable. Because at the heart of both of those things is an LLM, and so if you’re influencing the LLM, teaching it who you are, who you’re for, how to present you, then that’s going to work for both of those things.
Rich: So if we come from an SEO background, or if we understand SEO, and we have been for the past five, ten, twenty-five years, been thinking about things like keyword research and density, how do we need to shift our thinking to more of this prompt-based world that we’re living in now?
Tom: I think it’s just engaging with this is a fresh problem. And I think it’s not a coincidence that a lot of the advice is oriented around SEO metrics, the old SEO funnel, where we assume we can capture a lot of traffic, and then on site we look to try and convert that, right? That’s why a lot of people bemoan the lack of traffic from AI.
Well, it’s not necessarily because it’s not making an impact. It’s just not sending that traffic, so you need to capture that in a different way. But the SEO industry was kind of handed this problem first, and so my sense is that they tried to understand and solve it within their kind of existing frameworks, which is perfectly logical, but it’s a completely new problem.
AI systems operate very differently to search. Search is a directory, AI is a thought partner, is the way that I like to think of it. So to search takes a query, tells you where to find the answer. AI does helps you literally work through how to solve that in a bespoke way for you.
And there’s also lots of differences between a keyword and a prompt, right? A keyword is condensed. It forces you, as a user, to consolidate your question, your query, your thoughts, into a very short keyword. It actually encourages you to be verbose, to apply criteria. It encourages you to really expand the complexity, which obviously expands the number of possible variations of ways that people can ask it.
And we did some really basic Napkin math to look at a single keyword such as “best CRM solution”. When you think about all the different job titles, personas, criteria, use cases that people can apply to that prompt, you end up with twenty-two and a half thousand different variations of prompt that could come from that one keyword. So that’s the issue, or that’s why I would encourage people to start looking at this with a fresh lens. Because the traditional content marketing playbook is you find a query, you create a piece of content for that query, and you try and match that up one-to-one, and basically the best summary of knowledge against that query wins. That was search in a nutshell.
The challenge now is in order to do that, you’re going to need to create twenty-two and a half thousand pieces of content just for that one query that you used to create. And that’s where the big problems come in because then you create all sorts of issues with consistency and quality across your website, right? As much as AI can create content very rapidly, that’s a lot of content.
Rich: It’s an interesting thing you raise that the first people who either chose to take this on or handed the job are basically SEOs. And I think it’s one of those situations where if all you have is a hammer, all you see is nails.
So for those of us who maybe need to rethink the way that we approach online visibility and driving traffic to our website, and maybe that’s not even the right question, but I guess one of the things that we start thinking about is if this is all true, then what are the steps that we should be taking or how do we improve our visibility and the chances that the AI will recommend our company? What are the steps that we should be thinking about today?
Tom: So there’s three things that I think I would touch on there. The first is to separate your AEO and your SEO strategies. There are commonalities between the two. Both really hinge on your ability to build up authority, so the kind of EEAT side of SEO strategy, that’s going to help you.
And that’s where people do often say they’re the same thing. But you could say that about any aspect of marketing. If you are treated as an authoritative brand, all of your marketing works. If you’re not, it doesn’t. That’s a kind of universal truth, right? So a lot of marketing comes back to those fundamentals. It doesn’t mean that those things are all the same. So you need to separate them.
There are obviously elements of overlap that you can find efficiency and know that you need to really prioritize. And they’re, generally speaking, functions that sit closer to each other than say performance marketing or something. But that’s, I think, the first bit is viewing the separate things and then think about for each one of those things, what metrics are we going to track?
The AEO industry has adopted metrics that are derived from SEO, so it’s still looking at how much traffic do we get from AI systems. One thing our study showed is that if you look at all of those responses that we analyzed across complex journeys, only 16% of them ever produced a brand citation. Which means that the two big metrics that we’ve adopted, citations and traffic, if you’re looking at citations, you’re capturing 16% of the impact that AI could be having.
If you’re looking at traffic, we’ll think about, okay, it’s the citation is 16%. Let’s say 10% of people actually click on that citation, so that’s 1.6%. So that’s when we talk about those metrics, show you the tip of the iceberg in terms of AI’s ultimate influence on your industry, on your buyers, on their decision-making. You’re seeing a tiny fraction. So think about what you want to measure.
One of the things we really propose is actually focusing a lot more on fit. So rather than looking at visibility, which there is value to that because you still do want to show up for those bottom-of-the-funnel queries. But if you can focus in on how the AI views your strengths and weaknesses, and how that actually aligns with what buyers want in your target segments, and this is where it goes back to marketing fundamentals. Who are you? Who are you for? If you’re clear on that, you understand what trade-offs your buyers are looking for.
No solution is perfect. Everyone has trade-offs. You can map those and map the AI’s perception of your trade-offs, and that’s a much more predictive way to understand how is it going to present you across all of those twenty-two and a half thousand different variations of prompts that people might enter rather than tracking five of them and hoping that it’s the same as the rest of them.
So there are two things. If you’ve got me another minute and realize I’m rattling on, it would be focus in on the biggest core determinant seems to be information gain, right? So again, I’ll come back to that point of the old role of content. High-intent query, summarize knowledge against that query, best summary wins. That’s search.
Now, AI summarizes knowledge for every individual buyer on a bespoke basis. So if all you’re doing in your content is summarizing knowledge, AI doesn’t need you, humans don’t need you. Your goal should be to produce information gain, which is we define as net new knowledge that doesn’t currently exist. There are different levels of that.
There’s interpretive gain, so we actually have a framework that we allow people to measure. So interpretive gain is a new slant on an existing idea. So that produces a certain amount of value that might allow AI to go and pick it up and incorporate it into its answer.
Empirical gain, actual new data, original research, literally what I’m doing here right now. And then conceptual gain. I like to think what I’m doing here right now, which is that out of that original research has come an actual new concept that can further people’s understanding of a category They’re the kinds of things that actually make you useful to humans above and beyond what they could get not on your website, just talking to ChatGPT without you, and encourage AI systems to actually pick you up and incorporate you into their answers when they do go searching for answers.
Rich: So if that’s the future, is the suggestion then that we all need to be doing big research projects and bringing new information forward so that the AI platforms will see us as thought leaders or industry leaders and position us?
For somebody who’s just running an HVAC company, and I know that’s not your area of expertise, you’re in B2B, SaaS, and so forth, but where does that leave them? What type of content could the average business owner or marketer be creating within their industry that might warrant the attention of these AI platforms?
Tom: Yes, I think it depends on the complexity of your buyer, of how complex you are as a purchase or how complex your category is.
The three examples that I like to use sometimes when I’m asked this question are, think of three purchases that a head of finance is going to make. So toothpaste. There’s a very, very short journey from I’m out of toothpaste to buying toothpaste, right? You go, you see what’s on offer. There’s not a lot of category level influence that Colgate or Sensodyne or whoever need to do. There’s not a lot of innovation. It’s largely commoditized.
Trainers. There’s a little bit more category shaping that they want to do. If you’re buying trainers, you might want to know, okay, eco-friendly, I have pain, I’m a runner versus a football player. Whatever the different criteria is, there’s a lot more depth to that purchase.
Billing platform, an awful lot. It’s twelve to eighteen months of trying to understand how to take payments, recognition, all of these things that go into what I assume is quite a boring category. And I’m not going to list more of them. The more depth you have to the category, the more your category is won, not by whether you’re visible at the point of purchase, the tip of the iceberg, but whether you’re influencing the way that problems are framed, the way that requirements are built, right?
The simplest example is Nike. Let’s take one in the middle, Nike versus Adidas. Is it who wins the ultimate ad? Or if one is really good at X and the other’s really good at Y, is it who shapes consumers’ preferences towards X or Y? It’s the first one, and that’s very well known in B2B.
So the more you’re looking to shape your category, the more you need to really think about doing original research that really shapes your category. It’s like thought leadership is no longer something that is just a question of what you put on LinkedIn. You have got to back it up now.
Rich: So you did all this research. What were some of the things that you found most surprising based on what your original assumptions were once you started looking at the data?
Tom: The first thing that stands out to me is just how rarely AI, considering we call it AI search – and I’m actually trying to train myself out of that – just how rarely it searches. So again, we ran queries: awareness, consideration, conversion. But go on, let’s play. Guess what percentage of conversion queries AI actually searched for an answer?
Rich: Well, now I’m afraid of guessing at all.
Tom: If you had to guess, what would you guess?
Rich: Maybe it would be like 25%.
Tom: That was a bit more. So in conversion, it was 48%. So half of the time. In consideration and awareness, guess.
Rich: Oh, okay.
Tom: Zero. So never. So given that we know that the bulk of the journey…
Rich: For people who are not watching the video, Tom set me up by raising his hand much higher. Which is a nice psychological hack, by the way. But anyways, so continue on. Zero percent of the time during those initial stages.
Tom: Yeah, I sold you short there. I can only apologize. I guess I’m never coming back on this podcast.
Okay, cool. Yeah, so zero, zero, 48% as you go further down the funnel. That’s what I was doing. I was doing the funnel. So that’s quite interesting because we call it AI search. At conversion stage, what we noticed in the behavior therefore was as it’s going through the funnel, it starts off very exploratory. It’s working from training data, and it doesn’t ever search for an answer. Because the best hypothesis we have for why, is it doesn’t think it needs to. The intent behind the query is exploratory, and so it explores.
At consideration, the intent behind the query is trade-offs and understanding, and so it helps do that, and it does that from memory. It never once searched for an answer. Only at the bottom, when the intent starts to become decisional, does it actually go looking for answers. And even then half of the time. So that’s where it senses higher stakes and it kind of de-risks that by going looking for answers.
So if your whole strategy is built around citations and treating it as a search channel, you can only optimize for half of the queries in the conversion stage. So that shows just the sheer amount of the journey that you’re missing. And that trend of intent matching is what we’ve labeled it, was something that I found really, really interesting.
Rich: So awareness and consideration or decision, or whatever the exact phrasing you used. So awareness is, and we’ll use your CRM example. Awareness is, I might say something like, “I need a way for my entire sales team to be able to track where they are in the sales process,” as an example. I might just ask that question of an LLM, and it doesn’t need to go and do research because it has that knowledge, or at least enough of a knowledge, for it to give me a thing, “Oh, probably a CRM might be in your interest.”
And then we get down into consideration where we start talking about trade-offs. I’m looking for one that is good for a small team, or I’m looking for one that can scale up, or that can connect to Quicken, or whatever it would be. And again, what your research, if I understand it correctly, is saying is even at this stage, it probably has enough in its existing memory that it’s still not going off to do a search. But then as soon as I say, “Find me the five best CRMs that match what we’ve discussed so far,” now it starts to go do research.
Would you say that’s kind of an example of how that might work and why those numbers are the way they are?
Tom: Yeah, exactly that. It seems to be matching very closely the intent, which makes sense because it’s just a very smart algorithm that is designed to help humans solve problems. So it’s matching the intent of the query.
And we could see that in the language that we picked up as well. We did some analysis on that. And at the top of the funnel, the language is very exploratory. It’s kind of thinking with you, and then as you go further down, it goes to a little bit less exploratory, a little bit more kind of comparative trade-offs. That’s why comparison content does work quite well still, because you’re directly helping to enable that. And then eventually it becomes very, very decisional and it becomes a lot firmer.
It also becomes a lot broader and less variable in the brands that it surfaces. So at the top of the funnel, it surfaces brands with very high variability. There’s not a lot of consistency in terms of the brands that influence it.
At the bottom of the funnel, generally speaking, it surfaces the same brands over and over again, and it surfaces a lot more of them. Which is also quite interesting from a measurement point of view because it actually implies that at the bottom, the job’s largely already done. The real battleground where you can win outsized visibility, in quotation marks, is further up when it’s not really decided who to lean on.
Rich: It sounds like AI is already shaping the decision-making process long before it ever starts recommending anybody. If that’s the case, what can we as marketers do to keep ourselves in that mix or even to shape the type of feedback AI might be providing in those first two layers of the sales funnel?
Tom: I think it comes back to good marketing around… So I would say communicate fit for what we call the ‘option pool’ and be really clear on who you are and who you’re for.
And I’ll, I’ll give a bit more detail on those. So what we noticed was again, top of the funnel, option pool, very broad. And then over time, it kind of converges in a direction by applying criteria. The biggest opportunity we firmly believe is not to optimize at the bottom, but actually to help influence the direction of convergence so that you’re consistently pushing more users towards what you’re good at.
At that point, we know that visibility is largely decided because there’s very low variability at that point. I think of it as like a spotlight on the street, right? What you want to do is not necessarily run around chasing the spotlight but actually go nearer to it. And then you’re going to find that a lot more consistently you’re lit up. So that would be the main thing. The way that you do that is through traditional brand and is through quality, research-led thought leadership that can actually shape the way that a category thinks and the way that humans and AI conceive of their requirements.
I’ll give you a really good example that I think helps to illustrate this, going back to my days as a sales rep. I often describe myself as a recovering sales rep. And so you know what an RFP is, a request for proposal. So when a big company needs to buy something, they send that out to the market and invite lots of different providers to fill in this horrible, deep spreadsheet with fifty rows of very badly formulated questions.
Generally speaking, there’s a known cliché almost in the sales world, which is if you just get sent an RFP, you’ve already lost because one of your competitors has probably been talking to them for a year, helping them to understand what they should be looking for in a CRM. That’s kind of the old world.
So the way that I like to think of it is that if all you’re doing is optimizing visibility at the point of conversion, then you’re filling in the RFPs, but you’re filling in your competitor’s RFP that they shaped. They said what matters most. They wrote the questions so that it lent into their strengths. You want to go back a step and think more about how you can do that, which is completely counterintuitive.
Rich: I mean, that’s good overview in terms of… and certainly having helped businesses write RFPs and receiving them cold, I completely agree that when you have a chance to shape the RFP, you are in the best position to win it.
So can you give us some concrete examples of what we might be able to do to shape the questions or shape the responses that AI is giving at those higher levels that then would create that pathway to them recommending us when they get down to decision-making?
Tom: Yeah. So it’s the same four levers that you have for SEO or for anything else, really. Are you producing good content and are you capturing reviews? They’re the two big things, alongside your positioning, that are going to influence the way that you’re presented. And that’ll also give you the tools to reposition how the AI views your market and views what’s important.
I would say the two things to think about are from a thought leadership perspective, you can be hyper-specific. So that’s one thing that AI content generation or content kind of repurposing does allow. It allows you to be really clear on your website, this is an article for a CMO that explains why a CMO should want this in a CRM.
You can create very bespoke, specific content like that, that communicates what the requirements should be. The other thing though that’s going to have a bigger impact is research. And that’s where I do think what we’ve seen so far is most brands looking at AI as an opportunity to cut headcount, and they’re getting more and more efficient. But what they are going to see is diminishing returns as everyone does the same thing.
And I think you’ll see actually headcount build back up as people realize that we need to have people do clever things still, in order to get customers. But to your question, yes, I would be building out more of a research function that allows you to consistently generate those level 2 and level 3 information gain pieces of content that challenge the market and challenge the existing way of thinking. So it’s the same as it was before. It’s just a higher bar and it’s more critical now.
Rich: And does it make a difference whether we’re sharing this content on our website or sharing this content on LinkedIn or Medium or some third-party platform?
Tom: Yes, it does. It’s not clear yet, to me at least, exactly what difference. I think the simplest answer is, do it. You have to do it all these days. That’s where AI is fantastic. It’s great at repurposing a really good quality piece of research into content that can be distributed across Reddit, LinkedIn, your blog, every other channel.
I like to think of it as you have three levers to influence AI, your positioning, your content, your reputation. If you think of those three points of a triangle, we don’t know exactly how the three are weighted. We know that at the moment, reputation is a very, very big driver. And that’s what I would call Reddit, LinkedIn, G2, all of those kind of off-site sources that you have less control over. You want to be leveraging all three of the levers. You want to try and ensure that they’re painting the most consistent picture possible.
Because what you do then is you remove all of the ambiguity and any space for interpretation. The more distance you have between what your reputation says you are, what your positioning says you are, and what your content says you are, the more gaps there is for other people to fill in or for the AI to just make up.
It’s not very clear, I wouldn’t have the data yet to say 80% of your effort should go towards getting reviews, your reputation, 15% towards your content. I think that’s not clear yet.
Rich: If there was one thing that you would recommend us do if we want to get started on this, what would that first step be?
Tom: I think the first step would be measure the right things. So if you don’t have a relatively accurate understanding of the impact that AI is having… In fact, I’ll go back a step. The first step is to understand how your buyers use AI. Because there are people who come to me and they say, “We want to be seen by ChatGPT.” It’s like, “Right, who do you sell to?” “Well, seventy-year-old retirees.” “Okay, do you really?” Like, it’s a buzzword.
One of the challenges that we have a little bit or some brands have is that the early adopters of AI are CEOs and investors, and that creates a lot of downward pressure to show up in AI, even for brands for whom it is a relatively small channel. So that’s actually the first step.
The second step is to understand what impact it is having on you. If we assume that you sell to a cohort or a demographic of people that are using it, try to talk to them and understand how do they use it, what are they asking it at each stage so that you understand how to prioritize it. Because it is still probably going to be less of a revenue driver than something like search, unless you’re really selling to that early adopter ICP. And then the third one is if you decide we do need to prioritize it, and I would say for most brands, you need to, to at least some degree focus in on the right things to measure based on how complex your sales cycle is or your category is.
If it’s very, very simple, do what the existing advice tells you to do. Measure citations. Your goal is traffic because you’re selling largely a commodity. The more complex you are, the less relevant that is for you, and the more you need to look at something a bit more complicated, and the less you are likely to be able to unfortunately just pick out one metric and say that is all we care about. Because it is a more complex beast than that.
Rich: Now, I feel like we’ve only scraped the surface of the research you did. If people want to learn more about you, more about your company, and more about the research, where can we send them online?
Tom: Yeah. So head to our website, demand-genius.com. You can read the full report there. It’s called Dark AI, and that goes into an awful lot of detail. But also, if you’d like to, feel free to follow me on LinkedIn, Tom Radnai, and I’m relentlessly posting on LinkedIn just like anyone else these days.
Rich: Awesome. And we’ll have those links in the show notes as well. Tom, thanks so much for your time today.
Tom: Thanks for having me, Rich.
Show Notes:
Tom Rudnai has researched how large language models influence buyer behavior to help marketing teams better understand AI visibility, brand positioning, and evolving customer journeys. Check out the research his team at Demand Genius is doing and connect with him on LinkedIn to follow his latest insights on AI search and marketing strategy.
Rich Brooks is the President of flyte new media, a web design & digital marketing agency in Portland, Maine, and founder of the Agents of Change. He’s passionate about helping small businesses grow online and has put his nearly 30 years of experience into the book, The Lead Machine: The Small Business Guide to Digital Marketing.