ChatGPT and Google differ on top brand in 35% of buying topics, Bluefish says – PPC Land

Media teams at Bluefish clients get up to 50 keywords or ChatGPT context hints per ad group, ready for bulk upload. Can platform automation rewrite them first?ChatGPT and Google put a different brand at the top of their AI-generated answers in 35% of 5,810 buying topics, according to research Bluefish cited on October 27, 2026, when the New York company extended its AI marketing software from organic brand monitoring into paid advertising. The first product of that move, Paid Media Briefs, became available the same day. According to Bluefish, it gives media teams ready-made recommendations on ad targeting, creative and landing pages for Google and ChatGPT, drawn from data on where a brand is missing from, misdescribed in or already recommended by AI answers.The research behind the headline number is described in a single sentence, both in the announcement and in a product post on Bluefish’s website. Neither names the three categories it covered, the Google product it queried or the dates on which the answers were collected.Bluefish sells software that checks how AI tools such as ChatGPT and Google’s AI answers talk about big brands, and on October 27 it started giving those brands advice on buying ads in the same places. Its argument is that an ad shown next to an AI answer lands differently depending on which rival brands that answer already recommends, and Bluefish says Google and ChatGPT name a different leading brand for about one in three shopping topics. Clients now get ready-to-upload ad suggestions for Google and ChatGPT, although Bluefish has not published prices, results or the details of its research.Bluefish calls its product the Agentic Marketing Platform, or AMP, and sells it to very large companies. The platform is organized around what the company describes as four pillars: AI Monitoring, AI Optimization, AI Orchestration and AI Measurement. Until October 27, those pillars dealt with organic AI visibility – how often, and in what terms, brands are named in unpaid AI answers – a practice Bluefish refers to as Answer Engine Optimization, or AEO.The expansion carries all four pillars into paid media. According to Bluefish, the platform “identifies where a brand is absent, misrepresented, or already recommended in AI-generated answers, directs paid media to those gaps and opportunities, and measures the overall impact.” Paid Media Briefs, sold as an extension of the Bluefish Optimization Suite, is the first piece. The company says the briefs carry “platform-specific ad targeting, creative, and landing page recommendations designed to improve paid media performance across Google and ChatGPT.” No performance data accompanies that claim.On its own website, Bluefish promotes the launch under a different name, Bluefish Agentic Ads. The announcement itself uses “Paid Media Briefs” for the product and “The Future of Agentic Ads” as a section heading.Bluefish frames the problem in terms of context. “In this emerging agentic funnel, effective advertising depends on understanding the context in which ads appear in AI, including which competitors are recommended and how products are positioned,” the announcement says. The company adds that “most paid media teams lack visibility into the organic AI context surrounding their paid placements.”The product post on Bluefish’s website, credited to the Bluefish Team, gives more detail than the announcement. Briefs sit in the platform under AI Optimization, in a Paid Media Briefs section, and are pre-generated: they appear when the page is opened rather than being requested one at a time. They can be filtered by product line, topic and ad surface.Each brief covers one ad group on one AI channel. According to Bluefish, it explains why the visibility gap exists and compares how AI responses describe the brand and its competitors, then supplies search themes, keywords, headlines, descriptions, contextual guidance and landing page suggestions. Briefs can be moved into the bulk upload templates of Google Ads and ChatGPT Ads, which means the recommendations still pass through the advertiser’s own accounts and the platforms’ own systems before anything runs. Nothing in either document suggests that Bluefish places or buys the media itself.The labels used to classify gaps differ between the two documents. The announcement speaks of brands that are absent, misrepresented or “already recommended.” The product post lists absent, misrepresented and losing ground to competitors, applied to product lines and buying topics. It gives no scoring method or severity scale for any of them.The worked example in the product post is hypothetical: a consumer electronics brand whose portable fan is consistently missing from AI answers receives headlines that stress the product’s verified battery life. No named client, test or result is attached to the launch.Google briefs cover two surfaces, AI Overviews and AI Mode, according to Bluefish. Google has not offered either as a standalone buy; ads reach them through existing campaign types, and the briefs are formatted accordingly, for Search campaigns using AI Max and for Performance Max.For Search with AI Max, a brief holds up to 50 search themes or keywords, up to 15 headlines of under 30 characters each, up to four descriptions of under 90 characters each, and landing pages chosen from what Bluefish calls influential first-party pages. For Performance Max, the limits are up to 50 search themes, up to 15 short headlines of under 30 characters, up to five long headlines of under 90 characters and up to four descriptions of under 90 characters.The 50-theme ceiling matches Google’s own limit. Google raised the number of search themes per Performance Max asset group from 25 to 50 in May 2025, after starting with 10 in 2022.Both campaign types hand much of the matching to Google’s automation, which is where the evidence gets thinner. When Smarter Ecommerce looked at more than 250 retail campaigns in November 2025, AI Max conversions came in at roughly 35% lower return on ad spend than traditional targeting, and a separate four-month test put AI Max at $100.37 per conversion against $43.97 for phrase match. Neither was a controlled study, but both suggest that a brief’s keywords and themes are one input among several that Google weighs.The two surfaces also differ in what they accept. When Google expanded AI Overview ads to 11 more countries in December 2025, its documentation said ads would not appear in AI Overviews for sensitive verticals including finance and healthcare. Bluefish lists financial services among the 15 verticals it serves, and American Express among its customers. The product post does not say how briefs handle categories in which Google does not serve ads inside AI answers.AI Mode is the newer of the two. Google formally announced Shopping ads in AI Mode on February 11, 2026, and by May a sponsored product card was appearing in the middle of AI Mode answers, marked only by a small “Sponsored” label, as Google said the surface had passed 1 billion monthly users.ChatGPT briefs look different because ChatGPT ads work differently. Instead of keywords, a brief provides up to 50 context hints, plus ad titles of up to 50 characters, ad descriptions and landing page suggestions.Context hints are the phrases OpenAI asks advertisers to set at ad group level. When OpenAI opened its Ads Manager to all US businesses with cost-per-click bidding on May 5, 2026, its ad selection system did not operate on exact-match keywords, and context hints guided matching without guaranteeing delivery in any particular kind of conversation. By August, the hints were carrying the entire targeting burden inside ChatGPT, with reporting still grouped into two device buckets, Mobile and Desktop.The creative side faces a further variable. On October 20, OpenAI told advertisers that creative enhancements would be switched on by default for new and existing self-serve campaigns at the end of October, allowing its system to rewrite headlines and descriptions, translate ad text, resize images and change destination URLs, and to run those versions without the advertiser approving each one. OpenAI’s email gave October 30 as the start date; its help page said October 29. Unless an advertiser opts out, the titles a Bluefish brief recommends may not be the titles that run. The product post does not mention the setting.The research claim appears in the announcement in this form: “New Bluefish Intelligence research found across 5,810 buying topics and 3 categories, different brands lead AI-generated answers on Google and ChatGPT in 35% of cases.” Bluefish draws a commercial conclusion from it: “Advertising that is relevant on one AI platform may be poorly aligned with the answers consumers receive on another.”By PPC Land’s calculation, 35% of 5,810 topics is roughly 2,030. The same figure means that in about 65% of topics, or roughly 3,780, both platforms put the same brand first, a reading the announcement does not offer.Several things that determine what the number means are missing. Bluefish has not said which three categories were tested. It has not said which Google product was queried: the briefs target AI Overviews and AI Mode, while the company’s own boilerplate names “ChatGPT and Google Gemini” as major platforms it covers, and those are three separate surfaces with separate answers. There is no collection window, no account of how many times each topic was asked, and no definition of what it means for a brand to “lead” an answer, whether that is the first brand named, the one recommended or the one mentioned most often. AI answers vary between runs of the same prompt, so the number of samples per topic matters.Other measurements point the same way, even though none can be compared directly. Semrush’s AI Visibility Index, covering 126 million US prompts from January to April 2026 and more than 1,200 brands in 22 verticals, found only 36 brands that sat in the top 100 most-mentioned list on every platform in every month across ChatGPT, Gemini, AI Mode and AI Overviews. In a smaller study of UK software, Recrawled put 1,608 questions to the same four products on a single day in September and found that nine of 40 companies were never named in any shortlist answer. YouTube appeared in 25% of AI Overview answers in that study and in none of 360 ChatGPT answers – an illustration of how differently the platforms source their responses.Bluefish cites two figures to show how fast AI advertising is growing. Neither is linked to a source document.The first is that US “spending on advertising adjacent to AI-powered search results is projected to reach $26.4 billion in 2026, according to EMARKETER.” When EMARKETER published its US AI Advertising Forecast in June, PPC Land reported a 2026 total of $32.03 billion for US spending in and around AI platforms, split into three categories: ads beside AI summaries such as AI Overviews, which was the largest; ads in conversational search such as AI Mode; and ads inside chatbots. That report did not give a separate dollar figure for the search-adjacent category, so the $26.4 billion cannot be checked against it. The same forecast put US chatbot advertising revenue at less than $1 billion for 2026.The second figure is that “ChatGPT Ads has reached approximately $1 billion in annualized revenue run rate.” That milestone dates from August 31, 2026, when the run rate was reported at $1 billion, a figure derived from roughly $83 million of monthly revenue multiplied by 12. OpenAI’s David Dugan, who leads its global ads solutions team, wrote that the business had passed the mark in under 200 days; PPC Land counted August 31 as day 203 after the US pilot began on February 9.The two numbers are not in conflict, even though one is above $1 billion and the other below it. A run rate annualizes a single month of worldwide revenue; EMARKETER’s estimate covers a full calendar year, US only. What the gap does show is how far apart AI advertising totals can land depending on what is counted.The ad supply is growing regardless of the definition. On October 21, OpenAI extended ChatGPT ads to 26 more countries, for 89 in total, and said more than 100,000 advertisers had run campaigns since launch. In August, Similarweb said 26% of ChatGPT responses carried a sponsored ad and that nearly 30% of ad-eligible AI Mode queries showed ads, although the ChatGPT figure had first appeared as a US desktop measurement for June.Bluefish’s two named spokespeople are its cofounders. Alex Sherman, the chief executive, argued that marketers’ organizational charts no longer match how AI presents brands. “The agentic funnel doesn’t distinguish between organic and paid marketing, and neither should the platforms marketers use to manage it,” he said. “For decades, marketers have organized around channels, with separate teams, tools, and strategies for organic and paid media. But AI is collapsing those boundaries. We’re expanding Bluefish to give enterprise marketers a unified way to orchestrate that experience end-to-end.”Jing Feng, cofounder and chief operating officer, described AI visibility data as an input to budgeting rather than a report card. “AI is fundamentally changing how marketing strategies are built and executed,” Feng said. “A brand’s AI signal is becoming a critical input into marketing decisions, from where to invest to which campaigns to prioritize and how to measure their impact. We’re building Bluefish to make that intelligence the foundation for orchestrating enterprise marketing across every team.”Neither quote addresses measurement. The announcement says the platform “measures the overall impact” of paid media directed at AI gaps, but the product post describes no method for paid results. Its only statement on tracking is that organic performance on Google and on ChatGPT is tracked separately. Bluefish has not published pricing or said whether the briefs cost extra.Google and ChatGPT are the starting point. According to Bluefish, the platform will extend to four more kinds of paid activity: sponsored publisher content, creator partnerships, user-generated content and retail media. No dates are given for any of them. The company is also “developing capabilities” to bring its data into existing enterprise workflows through MCP and API integrations, again without a timeline.The publisher and creator items reflect how much of what AI tools cite is already commercial. Evertune’s chief executive, Brian Stempeck, found that more than 40% of the top 10,000 AI-cited sources contain affiliate links or are sponsored, a finding published when impact.com partnered with and invested in Evertune in December 2025. When Later launched a creator-focused AEO product in May, it said only 10% of AI search references come from a brand’s own site. If paid publisher content and creator posts feed the organic answers that Bluefish measures, the line between paid and organic that Sherman says AI is collapsing becomes harder to draw in the data as well. Scrunch’s September study of self-promoting listicles – articles in which brands rank themselves first – found that the tactic lifted the author brand’s recommendation rate from about 4% to roughly 7% when the listicle was cited.Retail media sits closest to the founders’ history. According to Bluefish’s Series A announcement in August 2025, Sherman co-founded PromoteIQ, the retail media software company Microsoft acquired in 2019. Andrei Dunca, the chief technology officer, co-founded LiveRail, the video advertising company Facebook acquired in 2014.Bluefish launched in 2024 and says it now works with “over 10% of the Fortune 500,” which would mean more than 50 of the companies on that list, with “hundreds of enterprise accounts spanning 15 verticals including travel, financial services, CPG, luxury and beauty.” The announcement names Adidas, American Express, eBay, Hearst and Tishman Speyer as customers. It does not say which, if any, have used the paid media product. In August 2025, the company said more than 80% of its customers came from the Fortune 500.The company raised a $20 million Series A in August 2025, led by NEA with Salesforce Ventures participating, which it said brought total funding to $24 million within 12 months of launch. In April 2026, it announced a $43 million Series B co-led by Threshold Ventures and NEA, with Amex Ventures, TIAA Ventures, Salesforce Ventures and Bloomberg Beta, and said total funding had reached $68 million. The two totals differ by $1 million from a straight addition, which Bluefish has not explained. American Express appears both as a named customer and, through Amex Ventures, as an investor.Bluefish is entering paid media from the organic side. Several companies have moved toward the same territory from different starting points over the past year.HubSpot released a $50-a-month AEO tool in April and said organic traffic for its customers was down 27% year over year. On October 5, Zeta Global agreed to buy Senso, the Toronto startup whose technology powers Zeta’s generative engine optimization offering. Microsoft Advertising published a retailer playbook in January on why products disappear from AI recommendations, arguing that filled-in product data fields drive ranking.From the paid side, Adthena launched AdBridge in April, which copies Google Ads campaigns into ChatGPT, and in May a ChatGPT Ads intelligence product that monitors more than 300,000 prompts a day. Similarweb added AI ad tracking across ChatGPT, AI Mode and AI Overviews in August.What distinguishes the Bluefish product, on its own description, is the link between the two: organic answer data is used to decide what paid media should say and where it should run. Whether that link improves results is the question its launch materials do not answer.Search advertising was built on a simple separation: organic results on one part of the page, ads on another, with different teams responsible for each. AI answers blur that arrangement. An ad in ChatGPT or Google’s AI Mode appears next to, or inside, an answer that may already recommend a competitor, and that answer differs between platforms. Bluefish’s 35% figure, whatever its method, puts a number on a gap paid search teams have not had to consider before: the editorial context of an ad placement is generated fresh by a model, and it is not the same on Google as on ChatGPT.The product also shows where control currently sits. Bluefish produces recommendations; Google’s AI Max and Performance Max and OpenAI’s creative enhancements then decide how much of them survives. A brief can propose up to 50 context hints and a set of ad titles, but OpenAI’s matching treats those hints as guidance rather than instructions, and its default-on rewriting can change the copy. The same is true of Google’s automated campaign types, whose performance independent testers have questioned.For enterprise budgets, the open questions are practical. Bluefish has not disclosed what the briefs cost, how it measures their effect on paid performance, or which categories its 35% finding covers. Until those are published, the launch says more about where AI visibility vendors expect advertising money to move – from tracking what AI says about brands toward paying to appear beside it – than about whether combining the two produces better results.Who: Bluefish, a New York company selling AI visibility software to large enterprises, led by cofounders Alex Sherman (chief executive) and Jing Feng (chief operating officer), with named customers including Adidas, American Express, eBay, Hearst and Tishman Speyer.What: Bluefish extended its Agentic Marketing Platform from organic AI answer monitoring into paid media and released Paid Media Briefs, which give media teams upload-ready keywords or context hints, headlines, descriptions and landing pages for Google AI Overviews, AI Mode and ChatGPT ads. It cited its own research finding that Google and ChatGPT name different leading brands in 35% of 5,810 buying topics, without disclosing the method.When: The announcement was made on October 27, 2026, with Paid Media Briefs available the same day. Sponsored publisher content, creator partnerships, user-generated content, retail media and MCP integrations are planned without dates.Where: The briefs cover Google, formatted for Search with AI Max and Performance Max, and ChatGPT Ads. Bluefish did not specify markets.Why: Bluefish argues that ads placed beside AI answers depend on what those answers already say about competitors, and that organic and paid teams should work from the same data. The company has not published pricing, results or a measurement method for the paid product, and platform automation on both Google and ChatGPT can alter what the briefs recommend.Every morning, get a curated digest of the last 24 hours in programmatic advertisingsource