Execution tools are commoditizing, but source behavior is not. In this solo news episode, Paris Childress covers Prefer's agentic execution product, Cloudflare's switchable web search API, Shopify opening agentic storefronts while Amazon and Kohl's block some agents, a Haus panel where YouTube dominates Google AI Overviews and AI Mode and barely appears elsewhere, 5WPR's KitchenAid engine split, Recrawl's finding that recognition is not a shortlist, citation head concentration, local GEO packaging off-site placements, GEO as a Clutch procurement category, and Search Engine Land framing AI brand visibility as bigger than SEO.
Execution is commoditizing, but source behavior is not. In this solo episode, Paris Childress, founder of Hop AI and co-founder of GEOforge, walks through this week's GEO stories.
In a Haus panel of 5,331 AI answers, YouTube appeared in 45% of Google AI Overview answers and 27.4% of AI Mode answers, versus about 1% in Gemini, 0.5% in Perplexity, and 0% across 1,164 ChatGPT answers. The panel covers 378 buyer questions from 16 accounts, not the whole web. YouTube is a strong lever for Google properties and a weak lever elsewhere.
Shopify agentic storefronts are active by default for eligible stores. The Wall Street Journal reports retailers splitting: Walmart, Best Buy, and Gap engage, while Amazon and Kohl's block some third-party agents. AI commerce visibility is increasingly about permission, catalog access, and checkout policy.
Recrawl's UK B2B shortlist study found nine of 40 companies never named in shortlist answers, even though every company was recognized when queried by name. 5WPR's kitchen appliances index shows KitchenAid strong in ChatGPT and Perplexity and weak in Gemini and Claude on the same prompts.
Subscribe to The GEO Show wherever you get your podcasts, and watch the full episode on YouTube: https://www.youtube.com/watch?v=U0biJ2_gH4c
Paris Childress [00:00]: Hi, everybody. Welcome to episode fifty-six of "The GEO Show," brought to you by GEOforge: Full Self-Driving for AI Visibility. And today is Monday, October fifth. We have some great stories to run through. The themes for today are that execution is commoditizing, but source behavior is not. All right, let's get right into the first story.
Paris Childress [00:32]: Prefer launches on Product Hunt as the execution layer for AI visibility. Prefer launched on Product Hunt on October third, ranking number three for the day. Its pitch is explicitly anti-dashboard. Understand how ChatGPT, Claude, Gemini, and Perplexity describe a brand, identify missing coverage, then use autonomous agents to act.
Paris Childress [00:59]: Prefer's own product materials describe an action center that connects tracked prompts and citations to prioritize the action and verifies results after changes ship. So yet another example of a new entrant coming right in at the execution layer. The visibility and read-only reporting layer is definitely not enough.
Paris Childress [01:23]: This has turned into a pure commodity, and I think the way that this category is evolving is that the execution layer is more and more dictated by agents who are making specific recommendations. Next story from Cloudflare. Cloudflare makes live web retrieval provider switchable. Cloudflare released its web search API in open beta on October second.
Paris Childress [01:57]: AI applications can switch among Ceramic.ai, Exa, and Linkup through one normalized interface. Search requests run through AI Gateway, which adds common logging, analytics, billing, access controls, zero data retention support, and verified bot requirements. So this is a pretty technical update here. Has to do with live web retrieval, which is the way that LLMs are basically fetching information to synthesize their answers in real time.
Paris Childress [02:35]: And, uh, Cloudflare is really at the center of that, the retrieval layer. So now the retrieval provider can become a configurable layer independent of the model. So two otherwise similar agents may see materially different evidence because they search through different retrieval infrastructure
Paris Childress [03:02]: Now on to a very interesting story about agentic commerce in the age of the massive rise of personal assistants like Meta Muse and now ChatGPT Dots and Instinct. Agentic commerce is becoming a permissioned distribution channel. Shopify documentation plus observed retail behavior. It has reported that Shopify says agentic storefronts are active by default for eligible stores, exposing products to ChatGPT, Google AI Mode and Gemini, Microsoft Copilot, and Meta.
Paris Childress [03:47]: With Shopify-managed settings, future agentic channels can also be enrolled automatically while direct checkout is supported on select services. The Wall Street Journal separately reports that retailers are diverging sharply. Some, including Shopify-powered merchants, Walmart, Best Buy, and Gap are engaging with agentic shopping, while Amazon and Kohl's are blocking some third-party agents.
Paris Childress [04:16]: Uh, this is a major story because this is really gonna dictate the future of how agents shop on the web. And certain providers like Shopify are opening up to agents because I believe that generally results in more incremental sales if you let agents come in and shop and transact on users' behalf. But then you have massive players, namely Amazon, who are saying, "No, uh, either you are going to use our agent," which is called Rufus, I think, "or we're going to block all the other agents."
Paris Childress [04:53]: So if you send your agent to Amazon to buy on your behalf, it will be blocked. You have to come yourself to buy. So it's a major development here. We'll see how it, how it plays out. Um, AI commerce visibility is increasingly governed by catalog access, merchant permission, checkout integration, and the platform's policy, not just public web optimization.
Paris Childress [05:23]: Next up, YouTube matters enormously to Google AI and almost nowhere else. Haus analyzed five thousand three hundred and thirty-one AI answers and forty-one thousand seven hundred and eleven source links across five surfaces. YouTube appeared in forty-five percent of Google AI Overview answers and twenty-seven point four percent of AI Mode answers versus just one percent for Gemini, point five percent for Perplexity, zero percent across eleven hundred and sixty-four ChatGPT answers.
Paris Childress [06:04]: The data set covers three hundred and seventy-eight buyer questions from sixteen Haus accounts, but it's not a web-wide consensus. Uh, so YouTube is getting more and more concentrated as a citation source, especially in Google AI Overviews and Google AI Mode, surprisingly not in Gemini and not in the other AI engines as well.
Paris Childress [06:32]: So it's very, very source-specific here. I still strongly advise people that want to move the needle on AI visibility for GEO, specifically in Google's properties, to definitely focus on YouTube and create more YouTube content. But this is not a universal GEO strategy because that YouTube content likely will not be cited very much or at all on the non-Google properties, and that includes Gemini
Paris Childress [07:06]: Next up, the same brand can be dominant in one engine and weak in another. And we've seen this before
Paris Childress [07:16]: It's confirmed that 5WPR, through its vendor and agency research published on October fourth, it ran a kitchen appliances AI visibility index with 98 prompts three times across four engines producing 1,176 answers. KitchenAid appeared in 59.2% of ChatGPT answers, 57% of Perplexity answers versus 22.8% in Gemini, 15% in Claude.
Paris Childress [07:49]: GE ranged from 45% in Perplexity to 18.7% in Claude. 5WPR explicitly notes that differences may reflect model retrieval design rather than an underlying universal brand score. So we've reported on this many times in this show. A blended AI share of voice can really conceal the surface where a brand has its biggest commercial weakness.
Paris Childress [08:17]: In this case, KitchenAid, a very well-known brand, was very strong in ChatGPT and Perplexity, and much weaker in Gemini and extremely weak in Claude for the same set of prompts. So it's very important to measure share of voice at the AI engine layer and not to blend it into an aggregate score.
Paris Childress [08:43]: Next up, being recognized by AI does not mean being recommended. Recrawl's AI shortlist study tested one thousand six hundred and eight questions across ChatGPT, Gemini, Google AI Overviews, and AI Mode. Nine of 40 deeply studied UK B2B software companies were never named in any of the 24 shortlist answers, even though every company was recognized at least once when queried by name.
Paris Childress [09:14]: The underlying experiment was run on September 4th, and the report was published on September 29th. So this is fresh coverage. So what does this mean? Uh, brand knowledge, brand mentions, and inclusion in a buyer shortlist are very different states. And of course, the goal here for brands is to make it to the shortlist.
Paris Childress [09:39]: But just because AI is recognizing your brand doesn't mean it's going to be mentioning you or putting your brand onto the shortlist. And we are seeing this in this study. Uh, nine out of 40 deeply studied UK B2B software companies were never named in any of the 24 shortlists, even though every one of those companies was recognized at least once when queried by name.
Paris Childress [10:08]: Okay On to the next story. Citation influence can concentrate in a surprisingly short source list In Cidea, vendor research analyzed two thousand six hundred and sixty-seven answers containing thirty-five thousand citations from a twenty-eight day B2B buyer question panel. In the subset where domains were readable, only twenty-nine of six hundred and twenty-three domains appeared in at least five percent of the answers.
Paris Childress [10:46]: While the top ten domains accounted for thirty-one percent of all domain citations. Source behavior varied sharply by engine. Perplexity averaged eighteen point six citations per answer, versus five for ChatGPT 4.1, and four point six for Claude. The data set is specific to HubSpot partner and growth agency questions.
Paris Childress [11:12]: So we're seeing in some categories, influencing a relatively small set of recurrent sources may matter more than indiscriminate outreach volume. And my advice here is that this is very, very, niche-dependent or industry-dependent. Um, sometimes there is a really long, long tail of citations, and in that case it makes sense to do a lot of outreach to a lot of diverse third-party sites.
Paris Childress [11:42]: And in other sectors, and in this case, this story is reporting a, uh, B2B buyer question panel. We don't know what specific industries, but it's B2B. There isn't much of a long tail. The citations are very, very concentrated at the head, and it would make more sense to focus your efforts on a smaller set of more dominant citation third-party sources.
Paris Childress [12:08]: Next story. Local GEO services are packaging off-site distribution as the product. Visibility360 announced an AI visibility audit and optimization offer for local businesses on October fourth. Its workflow researches category questions, produces news, video, podcast, and infographic content, and promises a minimum of three hundred media placements per campaign.
Paris Childress [12:36]: The placement guarantee and performance framing are company claims rather than independently validated outcomes. So we're seeing now a high volume off-site authority building coming from a local GEO service provider, and this is already being commoditized into a packaged service. So, uh, I'm very curious as to how this works.
Paris Childress [13:00]: They're guaranteeing per campaign three hundred media placements. So they are producing um, news, videos, podcasts, and infographics and pushing all that out to third parties for local geo visibility. Very interesting stuff The next story comes from Clutch. GEO is now an explicit procurement category on Clutch.
Paris Childress [13:33]: Clutch's dedicated Generative Engine Optimization company's marketplace shows ratings updated October fifth and describes GEO as also known as generative engine optimization or LLM optimization. Providers can be compared through reviews-- verified reviews, budget, industry, and service filters. So finally, GEO is a legitimate category on Clutch.
Paris Childress [14:00]: Took them quite a long time here, and just proof of the maturation of this, this new service category. GEO is no longer just a methodology debate. It is becoming a normal service buying category, and Clutch is probably the largest marketplace for, uh, B2B service vendors. It's the largest directory of B2B service vendors on the web as far as I know.
Paris Childress [14:30]: And I checked out the GEO company marketplace today, and there are already more than fifteen hundred agency providers of GEO services
Paris Childress [14:44]: And our last story of the day, reported by Search Engine Land. Search Engine Land's SMX, a major conference, is running a paid AI brand visibility master class today, October fifth, covering how AI evaluates brands, recommendation success and failure, organizational design, content, product, and measurement.
Paris Childress [15:09]: The workshop explicitly positions AI visibility as a cross-functional strategic problem rather than simply a new SEO tactic. And I really love hearing about this. Um, buyer education is moving beyond tactical GEO and toward organizational capability and operating model change. And I think this is an important distinction that now even Search Engine Land is making, is that GEO is much bigger than SEO.
Paris Childress [15:41]: GEO is not simply a new SEO tactic. It's not SEO two point oh. It's not a layer that you, that you put on top of SEO. It's much bigger. It has much bigger implications for all of marketing, and that is why it involves organizational capabilities, cross-functional organizational capabilities. And I believe already that GEO is migrating very quickly into brand marketing, and it is encompassing all of performance, such as SEO.
Paris Childress [16:16]: It's encompassing content and creative distribution, social, influencer, PR, and it's basically eating all of the adjacent disciplines of marketing. And it will become really AI marketing. I mean, I think that GEO effectively will become marketing in the AI age, and it's great to see Search Engine Land also giving some recognition to this as well.
Paris Childress [16:48]: All right. That's it for today. And, uh, from tomorrow, I'll be traveling to the How to Web Conference in Bucharest to talk about GEOforge. And then after that, I'll be headed to Brighton SEO, and I'll be speaking on Thursday morning, October eighth, about information gain at Brighton SEO. So I may not be able to keep rolling out these episodes daily for the next several days, but I will do my best.
Paris Childress [17:20]: So thank you all for listening, and we'll see you on the next one.
