Latest / Found in AI: AI Search Visibility, SEO, & GEO / OpenAI Is Turning ChatGPT Ads Into Conversations

Transcript

OpenAI Is Turning ChatGPT Ads Into Conversations · Found in AI: AI Search Visibility, SEO, & GEO

• 0:41
  SPEAKER_00: Hey, welcome back to Found in AI, the podcast that helps marketers understand what's actually changing an AI Search generative engine optimization and the way people discover brands online.
• 0:53
  SPEAKER_00: I am Cassie Clark, an AI Search Visibility Consultant, and today is September 17th, 2026.
• 1:00
  SPEAKER_00: Today we need to talk about what happens after your brand gets discovered in AI Search because OpenAI is turning ads into conversations.
• 1:09
  SPEAKER_00: Google is also giving retailers more tools to actually sell products inside those AI experiences.
• 1:16
  SPEAKER_00: And at the same time, Google is testing a program that pays publishers based on the value their content contributes to AI-generated answers.
• 1:26
  SPEAKER_00: As of now, that's really just invite only, but we have some information on it.
• 1:31
  SPEAKER_00: So let's just get into it.
• 2:01
  SPEAKER_00: So let's use the running shoe example, mostly because I'm a runner and I love talking about things like that.
• 2:08
  SPEAKER_00: So let's say that you are shopping for a new pair of running shoes.
• 2:12
  SPEAKER_00: You see the sponsored result, but instead of just clicking the ad, you could start talking to that brand's agent.
• 2:19
  SPEAKER_00: You could ask about sizing, or maybe even which shoe would work best for trail running, or whatever else you need to know before you are ready to buy.
• 2:28
  SPEAKER_00: And then, when you are ready to take the next step, the agent can send you to the advertiser's website.
• 2:35
  SPEAKER_00: OpenAI says these sponsored conversations will be clearly labeled and separate from both Chat GPT's independent answers and your original Chat GPT conversation.
• 2:47
  SPEAKER_00: For now, sponsored agents are being tested with just select advertisers here in the United States, but I think the bigger story is what this does to the traditional advertising funnel.
• 2:59
  SPEAKER_00: Hang with me for a second.
• 3:00
  SPEAKER_00: For years, the basic model has been relatively straightforward.
• 3:05
  SPEAKER_00: Someone searches for something, they see an ad, they click on that ad, they go to a website, and then it becomes the website's job to answer enough questions and overcome enough objections to get that person to actually convert.
• 3:19
  SPEAKER_00: OpenAI is potentially inserting an AI agent directly into the middle of that journey.
• 3:25
  SPEAKER_00: And it's not just changing the consumer-facing side of this.
• 3:29
  SPEAKER_00: OpenAI is also bringing more of the advertising workflow itself directly into ChatGPT.
• 3:36
  SPEAKER_00: Advertisers will be able to use natural language prompts in ChatGPT work to create, update, and analyze campaigns through an ads manager plugin.
• 3:47
  SPEAKER_00: OpenAI is also adding HubSpot and Shopify integrations to its advertising stack as well.
• 3:53
  SPEAKER_00: So this is not just a new ad format.
• 3:57
  SPEAKER_00: It's another example of something I've been talking about quite a bit lately, which is the difference between visibility readiness and agent readiness.
• 4:06
  SPEAKER_00: Because getting recommended by ChatGPT is just one problem.
• 4:11
  SPEAKER_00: But what happens when ChatGPT becomes part of the actual buying process?
• 4:16
  SPEAKER_00: That is a very, very different problem than just ranking for a keyword or showing up for a prompt because your content helps you do that.
• 4:26
  SPEAKER_00: OpenAI isn't the only company heading in that direction either.
• 4:29
  SPEAKER_00: Let's talk about Google.
• 4:31
  SPEAKER_00: So Google also announced another expansion of its Agentic Commerce tools this week.
• 4:38
  SPEAKER_00: There are a few pieces to this, so I'm just gonna go straight through them all.
• 4:42
  SPEAKER_00: Google is expanding its business agent into YouTube ads, which means someone watching a product ad can potentially ask questions there about that product without even leaving YouTube.
• 4:53
  SPEAKER_00: It's also continuing to expand Universal Commerce Protocol Power Checkout capabilities, which includes cart transfer and improved checkout testing.
• 5:04
  SPEAKER_00: And Google has also made AI performance insights in Merchant Center, generally available in several markets, including the US.
• 5:13
  SPEAKER_00: Those insights allow retailers to compare their share voice against other brands across AI mode and AI overviews.
• 5:21
  SPEAKER_00: Yep.
• 5:22
  SPEAKER_00: Uh-huh.
• 5:22
  SPEAKER_00: I hope you noticed that last part that I said.
• 5:25
  SPEAKER_00: Because retailers are now getting share of voice information for Google's AI experiences while the rest of us are still trying to pry useful AI visibility data out of Search Console.
• 5:38
  SPEAKER_00: We're gonna go back to that problem in just a minute.
• 5:42
  SPEAKER_00: But first, look at what Google and OpenAI are doing side by side.
• 5:46
  SPEAKER_00: OpenAI is testing how someone sees an ad, how they talk to an agent, how they ask questions, how they potentially buy.
• 5:54
  SPEAKER_00: Google is also building a thing over there to help people discover a product, help them talk to an agent, help them ask questions, and also potentially buy.
• 6:05
  SPEAKER_00: The interface is becoming part of the customer journey itself.
• 6:08
  SPEAKER_00: And I think that matters because a lot of marketers and brands are still spending quite a bit of time asking, how do I get my brand to appear in AI Search?
• 6:17
  SPEAKER_00: That is absolutely still an important question.
• 6:20
  SPEAKER_00: It is still the first step for all of this.
• 6:24
  SPEAKER_00: But, and I've mentioned this a couple times on the show recently, now there's a second question sitting right behind that one.
• 6:31
  SPEAKER_00: What can the AI actually do once it finds me?
• 6:34
  SPEAKER_00: AI visibility isn't necessarily the end goal here, it might just be the entry point to these new shopping experiences.
• 6:41
  SPEAKER_00: Now let's talk about a story that has had me scratching my head for a couple of days now.
• 6:47
  SPEAKER_00: And if you follow me on LinkedIn, you may have already heard me rant about this one.
• 6:52
  SPEAKER_00: I'm sorry, I have to bring it up on the podcast.
• 6:55
  SPEAKER_00: Google has, um, well, for the sake of sounding like an AI chatbot, um, but also it's really the only appropriate word here, Google has been quietly expanding what Digiday describes as a pay-per-value AI licensing program for publishers.
• 7:12
  SPEAKER_00: The pilot reportedly pays participating publishers when Google determines that their content has significantly contributed to an AI-generated response across Gemini, the AI overviews, or AI mode.
• 7:25
  SPEAKER_00: Participating publishers get an additional AI earnings widget inside of Search Console, which shows their monthly payout.
• 7:32
  SPEAKER_00: But, and here's where things get interesting, they apparently don't get much detail about how Google actually calculated that number.
• 7:40
  SPEAKER_00: Now, Diday reported on this, and according to them, the payments aren't based slippy on how frequently the content was used, but it's based on value.
• 7:50
  SPEAKER_00: So, in other words, Google is apparently attempting to determine whether a publisher's content meaningfully contributed to an AI-generated response, and then assigned some kind of monetary value to that contribution.
• 8:03
  SPEAKER_00: Now, if you know me, I have some questions.
• 8:07
  SPEAKER_00: Because Google has also been telling us how incredibly difficult AI search measurement is.
• 8:14
  SPEAKER_00: Earlier this week, it could have been late last week at this point, I'm not really sure, my days are running together.
• 8:19
  SPEAKER_00: John Mueller talked about how difficult it is to create useful position reporting for AI Search specifically, because the traditional idea of position one, two, three, four doesn't translate very neatly into those experiences.
• 8:34
  SPEAKER_00: Which I guess is fair enough.
• 8:36
  SPEAKER_00: I have said the same thing.
• 8:38
  SPEAKER_00: AI Search is probabilistic and measurement is messy, but messy is not the same thing as impossible.
• 8:45
  SPEAKER_00: And now we're learning that Google apparently has enough information internally to say, hey, this publisher's content contributed to this AI experience, that contribution was significant, and that's worth some money.
• 9:00
  SPEAKER_00: So it leaves me asking, Google, what exactly are you measuring?
• 9:05
  SPEAKER_00: Now, to be clear, those things aren't really a contradictory of each other.
• 9:10
  SPEAKER_00: Google obviously has access to internal data that publishers and websites don't have access to.
• 9:17
  SPEAKER_00: And the program is still an early pilot.
• 9:20
  SPEAKER_00: Digiday reports that even participating publishers describe the system as something of a black box.
• 9:26
  SPEAKER_00: Google pretty much says it's still learning how to best reward high-quality content.
• 9:32
  SPEAKER_00: And I think that is exactly why all of this matters.
• 9:35
  SPEAKER_00: Because publishers, website owners, marketers, all of us, we've been asking for better visibility into how content is being used inside of Google AI experiences.
• 9:47
  SPEAKER_00: And if Google can determine contribution well enough to attach a dollar amount to it, that really tells us something.
• 9:54
  SPEAKER_00: The measurement capability inside Google appears, appears, to be considerably more sophisticated than the measurement capability Google is currently exposing to publishers.
• 10:06
  SPEAKER_00: I am indeed talking about that AI Search Performance Report in Google Search Console.
• 10:12
  SPEAKER_00: So, we are starting to see the beginning of more sophisticated AI measurement emerge.
• 10:18
  SPEAKER_00: We might not have access to it all yet, and it's not necessarily showing up in the places where publishers or marketers have been asking for it either.
• 10:27
  SPEAKER_00: So those are the three stories that I have for you this week.
• 10:30
  SPEAKER_00: And the biggest question is what should marketers and brands take from all of this today?
• 10:36
  SPEAKER_00: As far as I can tell, based on what we just covered and what we know about these platforms so far, they don't want to just help people discover brands.
• 10:44
  SPEAKER_00: They're building more of what happens after discovery directly into that experience.
• 10:49
  SPEAKER_00: We're talking about asking questions, comparing products, making purchases, and even measuring the value of content that helped generate that answer in the first place.
• 10:59
  SPEAKER_00: Visibility still absolutely matters.
• 11:02
  SPEAKER_00: Again, that is the first step for all of us.
• 11:05
  SPEAKER_00: But we also need to be thinking about what happens once an AI engine actually finds your brand.
• 11:10
  SPEAKER_00: I said it last week, I will say it again probably a hundred more times.
• 11:14
  SPEAKER_00: That's the next layer of AI readiness that marketers and brands need to start thinking about now.
• 11:20
  SPEAKER_00: If you want this kind of analysis, straight to your inbox, subscribe to the Visibility Report.
• 11:26
  SPEAKER_00: Every Friday I break down what's changing in AI search and what marketers actually need to know about it.
• 11:31
  SPEAKER_00: It's a little bit more nerdier than what the podcast has.
• 11:34
  SPEAKER_00: The link is in the show notes, you can find it there.
• 11:37
  SPEAKER_00: I will see you all on Tuesday.
• 11:39
  SPEAKER_00: Until then, stay visible.