Stacking Growth · 2026-08-26 · 23 min
Key moments - from our scoring
Substance score
44 / 100
Five dimensions, 20 points each
B2B buyers increasingly rely on AI tools like ChatGPT and Gemini to research solutions, making traditional SEO rankings less valuable than being cited within AI-generated answers. Steph Pignola distinguishes SEO (traffic optimization via rankings) from GEO (citation and representation in AI responses), arguing they require fundamentally different strategies. AI prioritizes credible, consistent, and well-structured content - favoring FAQs, tables, and question-based headings over dense paragraphs. The episode examines why consistency matters (AI gets confused when brand descriptions vary across G2, LinkedIn, and your website), why gated content disappears from AI awareness entirely, and how AI struggles with audio, video, tabbed content, and metrics embedded in images. Real stakes matter: Chegg's stock dropped 40% after ChatGPT eroded their traffic. Practical tactics include ungating white papers, making transcripts accessible, writing opening sentences as if they're your only chance to be quoted, and using question-based subheadings to align with buyer queries. Measurement remains immature - no AI search console exists yet - so manual tracking across ChatGPT, Gemini, Claude, and Perplexity, plus analyzing form responses about discovery source, currently provide the best signals.
89% of B2B buyers use generative AI in their purchasing process, and AI overviews appear in 74% of problem-solving queries. If your company isn't cited in the AI's synthesized answer, you don't appear in consideration at all - there's no page of results to fall back on like there is with traditional search.
AI treats inconsistency as confusion and either leaves you out of responses or hallucinates inaccurate information about your company. Standardizing your name, description, and positioning language across all public sources is critical for AI to cite you correctly.
AI crawlers cannot read content behind forms, so if your best material is gated, AI has never encountered it and cannot cite it to buyers. Ungating your core content makes it visible to AI models and increases the chance it gets referenced in generated responses.
Lead with your point in the first sentence of every section (AI typically extracts only the first 1-2 sentences), use question-based subheadings that match buyer queries, and put key metrics and information in text, not images or hidden tabs - AI cannot reliably extract these.
Manually ask AI tools the questions your buyers would ask and track results in a spreadsheet by date, query, and tool (ChatGPT, Gemini, Claude, Perplexity differ in sources). Also add a simple 'How did you hear about us?' field to your demo request form to capture when customers cite AI as their discovery source.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode delivers a handful of genuinely useful tactical points - leading with the first sentence for AI extraction, putting metrics in text rather than images, question-based subheads - but they are padded with extended preamble, self-commentary, and well-known observations. The ratio of actionable insight to filler is moderate, not dense.
write the first sentence of every section as if it's the only sentence AI will ever read. Because functionally that is often true.
metrics inside of images...AI is not me. Uh, it is a robot. So you actually have to put that number in the text somewhere in the copy on the page
The GEO framing is real but already circulating widely in marketing discourse; the episode synthesises existing thinking competently rather than arguing anything contrarian or first-principles. The four-role framework (audience, analyst, gatekeeper, channel) is a useful organising lens but not a fresh invention.
AI isn't just a faster Google. It's become a bunch of different things together to a buyer all at once. It's the audience because if AI doesn't understand your message, it cannot repeat it.
SEO got you ranked, it doesn't get you cited.
This is a solo host monologue with no external guest whatsoever. The host is a practitioner at Refine Labs, a B2B marketing agency, which gives some practitioner credibility, but there is no evidence in the transcript of operating at exceptional scale, and the solo format eliminates any outside perspective or challenge.
I'm Steph Pignola
we at Refine Labs have been very, very bullish on this for several years
The episode cites several real, sourced data points (Forrester 89%, Search Engine Journal 74%, Chegg 40% stock drop, 56 meanings for MRO) and names specific AI infrastructure differences across tools. However, several claims lack traceable sourcing and the Chegg case is not analysed deeply enough to extract strategic lessons.
Chegg is an education company, publicly traded 2023. Their stock dropped over 40% after they told investors in their own words that ChatGPT was, quote, crumbling, unquote, their growth.
AI overviews now show up in 74% of what they call problem solving queries
As a solo monologue there is no host-guest dynamic to evaluate, which structurally caps this dimension; the host does anticipate listener objections reasonably well and maintains a light, accessible tone, but there is significant throat-clearing, self-interruption, and filler that dilutes the craft.
Quick question before we get into anything else. When's the last time you clicked through 10 blue links to actually research something? I'll wait.
Okay, um, let's talk about metrics, because I know you've been waiting 15 minutes and 30 seconds for me to shut up and start talking about metrics.
Computed from the transcript - who did the talking, and the words that came up most.
Generative Engine Optimization (GEO) is the practice of structuring B2B content so AI tools like ChatGPT, Gemini, Claude, and Perplexity cite your brand by name, instead of just ranking your website in Google. This episode breaks down David Kirkdorffer's incredible playbook on why traditional SEO alone no longer guarantees visibility, now that 89% of B2B buyers use generative AI somewhere in their purchase process (Forrester) and Google's AI Overviews appear in 74% of problem-solving search queries (Search Engine Journal).
Transcribed and scored by The B2B Podcast Index.
Speaker A: Quick question before we get into anything else. When's the last time you clicked through 10 blue links to actually research something? I'll wait. I bet for most of you it's been months, if not a year. You skim the AI overview and call it a day. That's basically the whole episode. That's the shift. I could stop right there and you'd have the headline. But I'm not going to stop there because the headline isn't the useful part. The useful part is what you can actually do about it. Stacking growth. So welcome back to Stacking Growth. I'm Steph Pignola, and today I want to talk about something we've been circling internally for a while. Which is what happens to B2B marketing when your buyer stops googling you and starts asking an AI about you instead? The uncomfortable part isn't that it's happening. Everyone knows that it's happening already. The uncomfortable part is that most of us built our entire content and SEO stack for a game that has been quietly changing under our feet. Let's ground this into something real instead of just vibes. I love vibes. But we want to get to the meat of this. There's a Forrester study out the B2B Buyer Adoption of Generative AI Survey from late 2024. 89% of buyers said that they were using Genai somewhere in their purchasing process. 89%, not just some buyers are experimenting. That's a lot of people. It's mostly everyone. And on the Google side, there's data from search engine journals showing that AI overviews now show up in 74% of what they call problem solving queries, which, if you think about it, is most B2B search. Nobody's typing CRM, um, into Google anymore. We've lost the encyclopedic keyword function. They're typing what CRM works best for a 40 person sales team that doesn't want to hire an ops person. And that's the other thing that's shifted. The questions themselves have gotten longer and much more specific. Voice search research shows the average query is something like 29 words now, which I think if you asked my middle and high school teachers, their brains might explode about. But that's where we are. It's what we're doing. People aren't searching. They're describing their situation and asking for an answer. And when they do that, there's no page of results to scan. There's one synthesized answer. You're either in it or you're not. Which is the Whole thesis. Honestly, SEO got you ranked, it doesn't get you cited. So let's draw out that distinction a little bit because I think a lot of people hear AI search and assume it's just SEO with extra steps. It's related, but it's not the same game. SEO is about ranking in a list. The goal is traffic to your site. Geo, which stands for Generative Engine Optimization, is more about being included as a source inside an AI generated response. The goal isn't traffic anymore, it's accurate representation and citation. Different goal, different metrics. Honestly, a completely different mindset, which usually is the hardest thing to change. And the metrics genuinely don't overlap that well between SEO, good SEO and good geo keyword rankings and organic traffic versus being mentioned by ChatGPT. It's like, oh, is he thinking about me? Uh, am I showing up in his lists? So how do you even track that today? We'll get to measurement a little bit later because that's the part that everyone asks about first and it's the part with the least, uh, amount of tools that have shown up, mature. But before we get into measurement, let's talk about why it matters beyond just, it's the new thing that's happening and we should care. So this is the thing that I don't want to be like an alarmist and say, you should be worried about this, but maybe you should worry a little bit, uh, about this. Uh, so it's not actually about AI itself. It's what happens to companies that are getting left out of the answer in those search queries. So there's a name brand example here worth knowing. Chegg is an education company, publicly traded 2023. Their stock dropped over 40% after they told investors in their own words that ChatGPT was, quote, crumbling, unquote, their growth. They've since sued Google, alleging AI overviews specifically hurt their traffic and revenue. I want to be careful here because that is an extreme case. And again, I don't want to be an alarmist. And the research definitely varies by industry. So, you know, your mileage may vary, but, but it definitely is an indication and a preview of what getting left out of the answer can cost. And I mean, I'd argue that it's not just about traffic either. It's what AI is actually doing in the buying process, which is a lot more than search. So I find this super interesting. AI isn't just a faster Google. It's become a bunch of different things together to a buyer all at once. It's the audience because if AI doesn't understand your message, it cannot repeat it. It's an analyst because it's curating and combining information the same way Gartner or Forrester, uh, could do. It's a gatekeeper, because if you're not in a generated answer, you're just not in consideration. And it's a channel because buyers are reading the AI summary and never actually clicking through to a website at all. The gatekeeper piece is the one that should maybe get you back to worrying a little bit. It's not that you rank lower, it's that you might just not exist in their shortlist at all. And it's not just research. That's the part that a lot of people are kind of stopping at. AI is expanding what buyers can actually do without ever talking to you. So there's an idea of zero click competitive intel. A buyer can just ask, how does company X, Y, Z compare to company abc? And get a whole synthesized answer pulled from your view. Sites, blogs, analyst pages, basically wherever is publicly accessible. If you haven't shaped that comparison narrative anywhere in the public, somebody else probably already has. And you might not be comparing as well as you'd like to, or the buyer just role plays with the AI, uh, act as a procurement lead evaluating solution X. What should I be worried about? If your messaging doesn't speak procurement's language anywhere in the open web, then AIA can't borrow that language and you'll never get into that conversation with that little role play. So the next question I think should be if AI is making these decisions about who gets mentioned, what is it actually looking for? Great question. So there's no dashboard, there's no universally agreed AI SEO score. But there are some patterns that we can dig into. Things that keep showing up across the research. Those are credibility, consistency and structure. So one at a time, credibility. AI models are prioritizing sources that are widely recognized research, reputable news analyst reports, well cited industry blogs. There's a study from Seer Interactive that found strong Google rankings actually still correlate pretty heavily with getting mentioned by AI. So SEO isn't dead. Even though we love saying things are dead on LinkedIn. It's just part of the input that isn't meant to say, throw out your SEO program. It's just that SEO alone isn't enough anymore. So that's credibility, consistency. This is the one that I think is most fixable and also maybe most ignored. AI notices when a brand shows up the same way across a lot of independent sources. So same name, same Description, same positioning language. If your homepage says one thing, your G2 profile says another, and a press release from two years ago says the third secret thing. AI treats that the way a confused buyer would. It gets uncertain. And when it's uncertain, it tends to just leave you out or make some things up. Hallucinate something that you don't want representing your actual company. So here's a little acronym example. Uh, if you just say something like MRO without spelling it out anywhere, that abbreviation alone has 56 different meanings, depending on which acronym dictionary you check. AI has no idea which one you mean unless you tell it clearly at least once. So you gotta be a lot more transparent and descriptive about things, maybe more than you used to. So the same logic applies to rebranding. We had a, um, B2B roundtable last summer with Dale Harrison who compared rebranding to a traumatic brain injury. And there's a lot, um, about that that kind of cycles into this AI conversation. Because if you are, uh, if you are rebranding over and over and over, if you rename your product, you're completely confusing the AI and it's not going to know how to represent you. So if you renamed your product, you should say formally old name pretty much everywhere for like a year. Because AI models won't actually know that company, uh, XYZ used to be called Company abc. So you've got to connect that dot pretty explicitly. So the third pillar that we talked about is structure. So structure is where it kind of gets a little techy, a little mechanical. AI favors content in FAQs, tables, bullet points, and clear headings because that's what's easiest to extract and reuse. If your key point is buried in the fourth sentence of a dense paragraph, AI, ah, might just miss it, or worse, quote it out of context. And then you're dealing with the same issue as before, that you are not represented in the way that you want to be or feel like you should be. So there's a specific tactic here that I think is really of interest. Um, and that is writing the first sentence of every section as if it's the only sentence AI will ever read. Because functionally that is often true. AI tends to lift the first one or two sentence, uh, and then that's it, that's, that's their, that's their entire context of you. Uh, so don't, don't do the thing where you build up to your point. You want to lead with it, say it, get it out of the way. So the primary benefits of X are abc and then elaborate after. Write like you want to be copy pasted because you actually do right now. I know it's like uh, something that we used to fear, but you actually want people to copy what you're writing now. Um, so there's a version for this for headings as well. Now stop calling a section Overview. Nobody's buyer question, nobody's asking overview. Call it how our platform handles data integration because that's actually answering something a person or an AI on behalf of a person might be asking. Question based subheads are underrated here as well. Again, don't just do section one or something like that. If a buyer might actually type how does pricing work? And your page has a subhead that says how does pricing work? Then that's about as close to a one to one match as you can engineer. So this is why we've done a lot of work in our YouTube videos this year, our podcast, our website, to reformat things into question types because we're trying to predict the questions that people are answer or people are asking so that we can be the ones that uh, these AI tools are pulling from, pulling the answers from. So I think a lot of people get surprised when they see good content, um, that is also completely invisible to AI right now. You can have the best videos, the best podcast, the best carousels on LinkedIn. You can have such good information out there and AI has no idea that you exist. Um, it's not because your content's bad, it's just because you haven't uh, repurpose the description or the title or the settings to be GEO first. So here's a little list, um, of what content you should specifically be looking at to change your GEO strategy. We're going to start with gated white papers and PDFs. Um, I know that we at Refine Labs have been very, very bullish on this for several years that you should stop putting things behind, ah, a form. Don't, don't, don't gate your content. Open it up. And that was because we genuinely believe that information should be shared and that people should be able to find the information that they're looking for without jumping through a billion hoops. But it actually kind of turns out that our perspective there was quite prescient for how AI has changed over the past couple of years. Especially because if it is behind a form, AI has never read it. And if AI has never read it, then AI is not gonna repeat it to anyone. So stop gating your content. Um, and then I wanna talk about something that is, like, sad to me, um, because it's been most of my job, which is webinars, recorded demos, podcasts, um, audio and video aren't machine readable without a transcript. So. So this is where we just have to change our strategy a little bit, right? I said earlier it's a mindset shift. It's not something that can happen overnight. This doesn't, this isn't a personal attack on me. It's not saying you like podcasts and you're bad because of it. It's saying that I just need to make sure that transcripts are easily accessible for that material. Now, YouTube and Spotify have been doing a lot of good work, um, with auto transcripts. So if that's not something that you have time for, that's, uh, you're, you're getting there. You're still gonna be starting to get visible, but you wanna make sure that those are accurate and, um, available so that the AI can pick them up. Uh, the third thing that I wanna talk about is anything living inside a tab, an accordion, or a carousel on your website, because a lot of crawlers only see what's loaded by default, not what shows up after a click. You wanna make sure that your most important information is just there. You don't want to have to expand it or click into it to know what it says. So that same thought applies also to metrics inside of images. A lot of us love doing. I'm guilty of this. A lot of us love doing, ooh, we had a 67% reduction in customer acquisition cost, and that's why you should hire us. And that's on, uh, a PNG or JPEG file. And I load the website and I would go, wow, that is 67%. That's. That's great. That's so cool. What a great company. Oh, my gosh, I should work with them. Um, AI is not me. Uh, it is a robot. So you actually have to put that number in the text somewhere in the copy on the page so that the AI crawler can read it. Um, what is exciting and might stand out to human eyes is not the same for robot eyes. So I hope that that reframe on your content is useful for you. Um, it's enabling AI to read your content and arm itself with all of the great things about your company that you want them to know or that you want potential future buyers who are asking questions to know. So we, you know, we said it before, stop gating everything. That doesn't necessarily mean that you have to Give everything away for free. Um, but you do have to honestly assess what you do want to put out for free. What do you want people looking in, asking questions in AI tools? What do you want them to know? And that's the stuff that you should ungate. That's the stuff that you should put out for free. So if it's locked behind a former tab, no one's going to know it. Okay, um, let's talk about metrics, because I know you've been waiting 15 minutes and 30 seconds for me to shut up and start talking about metrics. How do I know if it's working? How do I measure it? There's no search console for AI yet. Uh, so this is, uh, this is where we get a little wiggly, a little vibey. Um, I've given you some tactical tools so far, so hopefully those are helpful, but unfortunately there is no, like, good. One clear answer. Um, but honestly, you are a person who can use all the same tools that your prospective buyers are going to be using. You can ask ChatGPT, Gemini copilot Claude Perplexity. You can ask all of these tools the questions that your buyers would ask. Um, you can even use these tools to run an analysis on what the most asked questions in your field are. So what are the top solutions for category X? Which vendors offer capability Y? Who are the top competitors to us by name? Um, and you can track it literally in a spreadsheet if you have to, or just keep, um, checking, like, go back to it month over month to see where your company is or is not showing up. You add the date, you add the question, what AI tool you checked and what came back. Maybe we'll do a little template for that. It's a pretty easy spreadsheet, but in case you all want something, a neat little branded template for you to track your AI searches, then, uh, maybe we'll do that for you. Um, and you want to run it across multiple tools. That's super important. You don't just want to run only ChatGPT because everything retrieves information differently. ChatGPT and Copilot both lean on Bing a lot. Uh, Gemini uses Google's own search infrastructure. Obviously, Claude has shown signs of leaning on Brave Search. Perplexity pulls from a mix of sources and, uh, does pretty well to show its citation. So that might be some helpful context for you to know where these things are coming from. Um, and that is a nice one to audit because you can manufacture, you can track manufacturing what you've ungated or what you've Put out there and see how that's showing up in the answers. So, um, if you show up in one and not the others, that's something missing. You really need to take that into consideration. Or if you show up in many but not one, same thing. You want to use the information that this is giving you and not just say, oh, Claude hates me, bummer. Ah, you want to use that and move forward with it. So that is a way that you can track where to focus next. Um, also, we've talked about this for a while, but you want to ask your customers where they're finding you. If you have a demo field, like, uh, request a demo form, fill on your website, add an open question, open text field. How did you hear about us? That's four words. How did you hear about. Wow, that's six words. That is six words that you can put on your form that will unlock a world of information for you. We scan through that all the time. We're always looking for whether ChatGPT showed up or Claude or something. A lot of times we'll see, um, LinkedIn show up or YouTube or just Google Search. And that's how we know that our content is working. That's also how we know that our content is working. For these AI tools, it's always a little party. We put a little emoji, react when someone, uh, cites ChatGPT as how they found us. Because it means that all of this work that we've been talking about that we are trying to share with you is actually working and worth it. So that's a way to measure, um, and it's, I'm sorry that it's not like a one fix or, or easy tool. Um, and I know the tools are coming, I know that they're going to keep growing, but for now, you just want to keep an eye on it. So if you're listening to this and you've got like, I don't know, 30 minutes this week to spend on reworking your GEO strategy, this is the, this is the top, top. Couple steps that I would suggest. First, pull up your homepage, your about page, and your top performing blog post and ask if the first sentence of each section actually answers the question a buyer would be asking. If not just rewrite that first sentence. You don't need to go into a whole website redesign. Um, even though we just did and our website looks awesome now. So if you haven't been to refinelabs.com in a while, you should do it. Evan Hughes. Absolutely crushed. Um, but look at Your website. Look at the first sentence of every section and rewrite that if it doesn't answer the question that you're trying to get across. Number two, go check for consistency. Does your G2 profile say the same thing that your homepage says? Does your LinkedIn company page match that? Does your YouTube about section match that? Um, pick your brand and product description and make it identical everywhere. People want to know who you are always. Not just in some different places, some different ways. Number three, go ask. Pick a couple AI tools. You can choose your favorite. I'm not going to tell you which ones to choose, but ask them. Uh, your. A question that maybe your best buyer would ask about your category today. See what comes back. That question is going to be different for everyone, but it's going to be up to you. So you need a baseline for when you track this AI search result. Um, you don't need to have a perfect answer right now, but you need to have some place that you're starting. So SEO got you ranked, GEO gets you cited, and the brands that figure out the difference in the next m, I don't know, week, do those things that I told you, um, and see what happens. But like, in the next year, honestly, it's going to be harder and harder and harder after that to change and keep showing up. These algorithms change every single day. Um, David Kirkdorfer did all of this research for us in a whole, a, uh, whole tool, a whole, uh, playbook for geo. And I know that even since he did this research, some of the, um, some of the crawlers have, have changed. Not in massive ways, but like just the, the places that they pull from. So make sure that you are staying on top of things and if you don't change it now, again, it's going to be harder to change in the future. So that's it, that's, that's the episode. I could have stopped, uh, after the first sentence, but I didn't. And now you have 22 minutes of me talking. Um, I hope it was useful. I hope you send it to whoever is owning SEO at your company. Not because you're trying to like, shame them or say that their job isn't important, but because you want to have this conversation with someone who is so in touch with how the brand shows up. Um, that's, that's it for me. I hope you all enjoyed. Let me know if you want more of this or if there are any questions that you have that you want us to touch on in a future episode. Okay, have fun. Bye.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.