OpenAI's ad push begins, and The Knot is co-piloting

Summary

---

The Digiday Podcast: OpenAI's Ad Push Begins, and The Knot is Co-Piloting

Date: March 3, 2026
Host(s): Kamiko McCoy (Senior Marketing Reporter, Digiday), Tim Peterson (Executive Editor, Digiday Media)
Guest: Jenny Lewis (CMO, The Knot Worldwide)

---

Episode Overview

This episode dives into OpenAI’s move into ad partnerships, spotlighting its first ad tech deal with Criteo, and features an in-depth interview with Jenny Lewis, CMO of The Knot Worldwide. The discussion explores The Knot’s role as an early ChatGPT advertising partner, the emerging landscape of “answer engine optimization” (AEO), and how brands can stay relevant as AI transforms both search and advertising. Jenny details The Knot's experimental foray into generative search, their ChatGPT app, and best practices for marketers in a rapidly changing environment.

---

Key Discussion Points & Insights

1. OpenAI’s Military, Ad Deals, and Industry Backdrop

• Pentagon & AI Ethics:

  • OpenAI recently inked a deal with the Pentagon after Anthropic declined due to ethical concerns, specifically around mass surveillance and human control over the use of force.
    • Tim Peterson [02:03]:

      "Anthropic said, 'We'll do a deal with you. We just have two stipulations. One, don't use our technology for mass surveillance. Two, if you're going to use our technology to kill people, we would like to make sure that a human is the one pressing the button.' And the Pentagon is like, nah, that's not going to work for us. And so Anthropic says, all right, no deal then."

  • There is skepticism about OpenAI’s claims of strong “guardrails” in their own deal, with reports indicating the Pentagon did not budge in its terms.
    • Tim Peterson [03:05]:

      "OpenAI agreed to follow laws that have allowed for mass surveillance in the past, while insisting they protect its red lines. So not great."

• OpenAI-Criteo Ad Tech Deal:

  • OpenAI announced its first ad tech partnership with Criteo, signaling that ChatGPT—with its massive 900 million weekly users—is officially scaling advertising.
    • Tim Peterson [04:24]:

      "Normally you do a deal with an ad tech vendor when you have so much supply that you need to fill it with additional demand."

• Audience Backlash:

  • #QuitChatGPT trended due to concerns about AI’s role in surveillance and ethics. However, this hasn’t slowed OpenAI’s commercial push.
    • Kamiko McCoy [04:03]:

      "The hashtag Quit ChatGPT is not surprising to me that that was kind of the response to all that's happened. But it hasn't stopped OpenAI from moving."

---

2. The Knot's Experience as a ChatGPT Pioneer

A. Becoming One of the First ChatGPT Advertisers

• OpenAI approached The Knot due to their prior experimentation with a ChatGPT-integrated app and willingness to test the platform.

  • Jenny Lewis [08:03]:

    "We were honored to be approached to be a part of the ADS pilot...I think because they knew we were tinkering in the space, they knew that we had interest."

• The partnership is firmly in the test-and-learn phase; even The Knot doesn’t yet know what the big advantage will be, but wants to jump in early.

  • Jenny Lewis [09:05]:

    "It's early days, so I don't know that we know yet what the advantage will look like, but we want to be on the front lines to figure that out."

B. Initial Partnership Logistics and Safety Concerns

• Top Considerations:

  1. Brand and user safety, including handling of user data.
  2. Clarity on data and analytics provided by OpenAI.
  3. True partnership, with regular checkpoints—not just a plug-and-play ad buy.
  • Jenny Lewis [11:27]:

    "I think for us, first and foremost was safety...Second, understanding the data that we would be getting back...and then third is just the level of partnership..."

• Deliverables:

  • The Knot provided ad copy and images, mainly adapted from previous campaigns, with copy customized in-house.
    • Jenny Lewis [15:54]:

      "We were able to pull most of the imagery just from, you know, shoots that we've done in the past...And then the copy we customized for through our in house copywriting team."

C. How ChatGPT Ads Look and Behave ([13:25–15:25])

• Appear as standard but visually integrated ad units beneath ChatGPT responses.
• The Knot is using ads to promote its Vendor Marketplace, matching couples with local florists, DJs, photographers, etc.—a behavior already organically happening within LLM-based searches.
  • Jenny Lewis [14:07]:

    "They have a visual that will do it much better justice than what I will attempt to describe...We focused our ad campaigns specifically on Vendor Marketplace..."

D. Data & Metrics

• The Knot began receiving performance insights about the ChatGPT tests very recently, with the lens on learning rather than expecting immediate performance. Data specifics remain confidential.
  • Jenny Lewis [17:03]:

    "We're still taking them in and kind of contextualizing them...Our assumption going into it...it's not that it's going to be like a fully optimized channel out of the gate."

E. Branding vs. Performance: The Knot’s Philosophy ([17:44])

• The Knot treats all campaigns as serving both brand and performance goals, refusing to rigidly separate them.

  • Jenny Lewis [17:44]:

    "Brand is performance and performance is brand...A consumer is not looking at an ad and saying like what a great brand marketing campaign or like, oh, there's a performance campaign, I must click on it."

• Early OpenAI ads are funded through the Knot's test-and-learn budget, a “slush fund” for experimental channels.

  • Jenny Lewis [19:24]:

    "We try and every year carve out a bit of a slush fund that's just reserved for new channels...This tapped into those dollars."

---

3. The Rise of Answer Engine Optimization (AEO) and GEO ([24:38–28:04])

• In Jan 2025, The Knot’s SEO team rebranded themselves as “Search Everywhere Optimization," including not just Google but LLM search engines and social search.

• Early lessons in AEO:

  1. Tech Fundamentals: Enable proper crawling/bot access for AI models without compromising proprietary data.
  2. Intent-driven Content: Focus content on answering direct questions—understand, via tools like Profound, what users actually ask.
  3. Content Distribution: Don't limit optimization to your own website—leverage forums, social platforms, reviews (Trustpilot), and press to establish authority and sentiment.

• The Knot's historical investment in search and content positioned the brand well for visibility in AI-driven environments.

  • Jenny Lewis [28:04]:

    "We have built such credibility and authority on Google over such a long period of time...So our visibility out the gate was reasonably strong."

• ChatGPT and similar platforms have changed content marketing from a “front door” (Google) into a world with “no walls":

  • Jenny Lewis [32:46]:

    "Google's the front door and now it's like you have a front door, a side door, a garage door, a back door."

  • Jenny Lewis [32:57]:

    "No walls. You're just in like the middle of a campsite..."

• Help centers and FAQ schemas play an increasingly important role as LLMs crawl and ingest answers to questions.

---

4. Marketer Takeaways & Budgeting Amid AI Search Disruption

• Budgeting:
  • The Knot’s shift toward social has been driven by consumer behavior rather than just AI search. There hasn’t been a dramatic overhaul in allocations specific to AI so far, but the brand ensures it has budget flexibility for emerging opportunities.
    • Jenny Lewis [34:57]:

      "I'd say as I look at our media budget and allocations overall, we've seen a bit more of a skew towards social than we have in the past...But I wouldn't say that was spurred by AI search necessarily more by consumer behavior at large."

---

5. The Knot’s ChatGPT App: Early Experiments ([36:02–39:05])

• The Knot released a ChatGPT app so users can pull vendor suggestions directly inside ChatGPT. Marketing around the app is intentionally light, allowing the team to observe organic adoption.

  • Jenny Lewis [37:23]:

    "It's new and like any good beta product, you're not going to do a massive go to market to promote it..."

  • Jenny Lewis [39:05]:

    "What we're really trying to most understand is user adoption and behavior, especially with such light marketing..."

• The team wants to learn how users interact natively and how the app funnels back to The Knot’s main site or app.

---

Notable Quotes & Memorable Moments

• On the brand’s test-and-learn philosophy:

  • Jenny Lewis [09:05]:

    "We want to be on the front lines to figure that out. And so hopefully this will allow us access to be able to do so and to provide feedback to help shape the product and help OpenAI figure out what's going to be most valuable for marketers and ultimately for customers."

• On AEO and the changing nature of content marketing:

  • Jenny Lewis [28:04]:

    "It is exactly as that executive said, you know, like ensuring, like doing some of the dirty work in understanding, like what are you allowing them to crawl? What do you not want them to crawl..."

  • Jenny Lewis [32:46]:

    "It's like, you know, they used to say, like, Google's the front door and now it's like you have a front door, a side door, a garage door, a back door."

• On search behavior and strategy:

  • Jenny Lewis [24:38]:

    “In probably January 2025, I actually rebranded our internal SEO team for the acronym to stand for Search Everywhere Optimization and broaden their aperture to think beyond just Google.”

• On data transparency challenges in AI search:

  • Jenny Lewis [22:57]:

    "It's hard to tell because of the way the data that we're getting is structured...It's not as straightforward as this ran in an AI spot or this ran in an organic search spot."

---

Important Timestamps

• 02:03 — Anthropic and OpenAI deals with the Pentagon; ethics and surveillance.
• 04:24 — OpenAI's first ad tech partnership with Criteo and implications for ad inventory.
• 07:29 — Start of Jenny Lewis (CMO, The Knot) interview.
• 08:03 — How The Knot was selected as a ChatGPT ad pilot partner.
• 11:27 — The Knot’s considerations: brand safety, data transparency, partnership structure.
• 13:25 — The first sightings of The Knot ads in ChatGPT and what they look like.
• 14:07 — Explanation of The Knot’s Vendor Marketplace ad focus.
• 17:44 — Discussion on whether ChatGPT ads are performance or brand marketing.
• 24:38 — The Knot’s rebrand of their SEO team to include AI and social search.
• 28:04 — AEO best practices, lessons learned, and importance of diversified digital footprint.
• 32:46 — Analogy of new search environment: “no walls” in digital marketing.
• 37:23 — The Knot’s ChatGPT app: strategy and early stage learnings.
• 39:05 — The Knot’s hopes for learning user behavior from their ChatGPT app.

---

Conclusion

This episode offers a rare window into the real-time, tactical decision-making by major brands confronted with the generative AI revolution in search and advertising. The Knot’s collaboration with OpenAI exemplifies a nimble, test-and-learn approach, while their experience in AEO gives practical guidance to other marketers. The future of brand building and performance in the age of AI will demand even more experimentation, flexibility, and a keen eye on where consumer attention is heading.

Host’s closing thought: "You all are definitely on the front lines. One of the first brands to advertise on ChatGPT, a brand that was thinking about AEO and GEO in January of 2025..."

---

Additional Resources

• Profound: Third-party tool mentioned for AEO insights
• Criteo: Ad tech partner for OpenAI’s ChatGPT monetization

(Note: Ads, intro/outro, newsletters not included.)

---

Transcript

[A]

Foreign.

[B]

Hello, hello, and welcome to another episode of the Digiday Podcast, a show for anyone who's still wondering, is ChatGPT getting paid to say this? I'm Kamika McCoy, senior marketing reporter here at Digiday.

[A]

And I'm Tim Peterson, executive editor, Video and audio, Digiday Media. What's up, Kimiko? How you doing, Tim?

[B]

Things are. Things are happening too fast, too quick, everywhere, all the time, all at once. How are you hanging in there?

[A]

I. I am happy to have a reprieve from talking to people about Paramount now coming back in and snatching Warner Brothers Discovery away from Netflix, because that has dominated my morning. It's only 9am My time here in California. I've already had three calls talking about, which obviously is a huge story because I think we were all expecting, like, okay, Netflix kind of seems like it has this in the bag. Warner Brothers Discovery does not seem to really want to entertain Paramount's offers. But Paramount also isn't going away. And then it got to the point where Warner Bros. Discovery said last week, actually, Paramount's latest offer, pretty good. That's a lot of money. Netflix, do you want a match? And Netflix is like, nah, I'm good. I'm instead going to take my almost $3 billion payout from Paramount for not acquiring you and walk away. Hope you two are happy together. And that's the world I live in. But again, I'm happy to step out of the world, out of that world to talk with you about ChatGPT. Because, I mean, that's another thing. There's so much going on right now with ChatGPT.

[B]

Yeah, absolutely. So on the one hand, you've got OpenAI, which has inked a deal with the Pentagon after Anthropic said, no, thank

[A]

you, After Anthropic said, we'll do a deal with you. We just have two stipulations. One, don't use our technology for mass surveillance. Two, if you're going to use our technology to kill people, we would like to make sure that a human is the one pressing the button and decide, like, pressing the kill button. And the Pentagon is like, nah, that's not going to work from us. And so Anthropic says, all right, no deal then. To which then Sam Altman and OpenAI come in and say, like, hey, we'd like to do a deal now. OpenAI is claiming that they have implemented the same. They've put the same guardrails in place with their Pentagon deal, which I don't think has been, like, formally signed or agreed to yet, but that like they have the same guardrails that Anthropic was supposed to have. But then this morning, the Verge reported, Hayden Field over at the Verge reported that actually, no, those claims are wrong. I'll quote from her story. The answer, sources told the Verge, is that the pentagon didn't budge. OpenAI agreed to follow laws that have allowed for mass surveillance in the past, while insisting they protect its red lines. So not great. Not great for OpenAI. It's actually, in some ways it's not great for Anthropic because of this supply chain designation that goes a bit beyond the scope for you and I. But that could affect Anthropic's dealings with Nvidia, Google and other companies that deal with the military. But over the weekend, Anthropic's Claude, its ChatGPT rival, became the number one app in the App Store. And at the same time, my Instagram threads feed on Saturday was full of posts of people posting about quitting ChatGPT.

[B]

Well, I think that goes back to like the whole fear that people have had around AI for a long time around mass surveillance and things like that. So the hashtag Quit ChatGPT is not surprising to me that that was kind of the response to all that's happened. But it hasn't stopped OpenAI from moving. They've just announced a deal this morning with Criteo.

[A]

Yep. So Critio is the first ad tech vendor that OpenAI has done a deal with, which actually surprised me when I first saw this announcement, because normally you do a deal with an ad tech vendor when you have so much supply that you need to fill it with additional demand. Now, I mean, OpenAI, I think last week announced that ChatGPT now has 900 million weekly users. That number may be a little bit different as we speak than what it was last week, but still, that's a ton of people. That's nearly a billion people using ChatGPT each week. And assuming that a large chunk of those people are in the free tier, those are the people who are going to see these ads. So there probably is a lot of inventory that ChatGPT could be filling or OpenAI could be filling in ChatGPT. So what that suggests to me then is Anyone on the ChatGPT free tier or the go tier who hasn't been seeing ads. That's probably going to change in a few weeks when this Critia deal takes effect.

[B]

Yeah, and that's if you haven't already seen ads from the Knot, which we talked to the Not CMO for this episode, to understand what the ChatGPT ads look like, what the direction that they're going here, and what it means for a company like the Knot.

[A]

Yeah, yeah. So we spoke with Jenny Lewis, who, as you mentioned, is the CMO of the Knot Worldwide, and I thought it was a really fascinating conversation. We spent a lot of time really understanding how the Knot came to be one of the first advertisers on ChatGPT, what the strategy, was there, some sense of how things are going. But then we get into how they think about ChatGPT in kind of like the broader context of. Is this part of the search play? And we get, like, really into that because the Knot has, you know, search has been an important channel for the Knot. And then we also get into AEO and geo. Oh, my. Oh, my.

[B]

Lions, tigers and bears.

[A]

Yeah, exactly. Because that's something that Jenny mentions in the conversation, how it was January 2025 when the knot really started digging in on answer engine optimization, Generative engine optimization, basically. How can the Knot make sure that it's getting mentioned in these AI conversations people are having? So this is a brand that was really early to AEO and Geo, and so therefore has, I think, a really helpful perspective on the matter as well as, Obviously, insight into ChatGPT advertising.

[B]

Absolutely. Listen, Jenny with a big J has got a good, juicy scoop for us, so I won't hold up any longer. We can go ahead and get to the conversation.

[A]

Jenny Lewis, welcome to the DJ podcast. Thanks for joining us.

[C]

Thanks so much for having me.

[A]

We're very excited to have you because the Knot is one of the first advertisers on ChatGPT. And, Kameka, we can't stop talking about ChatGPT ads. I feel like it's been the dominant topic on the show for the past six months at least.

[B]

Yeah, I think that's a fair timeframe for us, if not more. Yeah.

[A]

So, Jenny, how did you all get approached. When did you get approached to be one of the first advertisers on ChatGPT?

[C]

Sure. Yeah. Thanks so much for having me. And I love that you all have been talking so much about ChatGPT on the podcast. We have definitely been talking about it quite a bit this past year. We were honored to be approached to be a part of the ADS pilot. It wasn't all that long ago that they reached out to gauge some interest. I think we'd been previously building a not app to live within ChatGPT and the first in the wedding industry. And I think because they knew we were tinkering in the space, they knew that we had interest, and so we were thrilled to jump on board and just learn some more.

[B]

I'd love to get a sense of, like, how that partnership came together. Right. Because, like, on the one hand, you're building out your own deal, but on the other hand, there's a lot of this bigger ecosystem that's being built out with. With. With OpenAI and ChatGPT and whatnot. Kind of talk to me a little bit about what the interest was, what the hope is, and having the advantage of using this ad platform.

[C]

Yeah. So I think all that said, it's early days, so I don't know that we know yet what the advantage will look like, but we want to be on the front lines to figure that out. And so hopefully this will allow us access to be able to do so and to provide feedback to help shape the product and help OpenAI figure out what's going to be most valuable for marketers and ultimately for customers. For us, it really has been all about from day one, making sure that we are where our consumers are. And that, of course, is marketing 101 in some ways. But for us, it's that much more important, given you're only engaged for, on average, about 18 months. So it's important that when you get engaged, the knot's top of mind. We're engaging you early in that wedding planning process so that we can be there throughout the whole full journey. And so for us, you know, what we saw over the past year is that about 36% of couples were using AI in the wedding planning process. And now that's not just ChatGPT, but of course, they're a big part of that pie. And so it's natural for us, as we think, in the context of our broader marketing strategy of be where the couples are, for us to want to be there, too. And so, again, early days don't know yet what exactly it's going to produce, but we're viewing it very much from a test and learn standpoint so that we can hopefully help figure out where this is going to fit into our marketing strategy in the future.

[A]

And you mentioned this was a quick turnaround for you all. Was it this year that you were approached by OpenAI?

[C]

Was it this year? I believe so. Gosh, what year is it? Even? I'm buried in snow in the Northeast right now. It's like, I don't even know what's going on, but I think it was early January. Yeah.

[A]

Okay. And so, like, what were the initial questions that like you and your team had in evaluating the opportunity because on the one hand, opportunity to be the first advertiser on a major platform like ChatGPT, very exciting. On the other hand, being the first advertiser on any platform means that platform is also still figuring these things out. OpenAI has presented this as a pilot, so implicitly is figuring things out from your end. What was it that you needed to figure out out what were the questions that you had as you're evaluating the opportunity here?

[C]

Yeah, I think for us, first and foremost was safety. Just brand safety, user safety, how we're thinking about how they intend for us to show up and understanding how data will be handled. Second I would say was understanding the data that we would be getting back from the platform to help evaluate is this effective or is it not? And then third is just the level of partnership and ensuring that we weren't just going to be slotted in and never hear from them again, but that we would have recurring touch bases, which is our expectation of all the ad platforms that we partner with, that we're doing QBRs, that we have shared business goals, especially if we're investing at meaningful levels with you. And that very much was the case from the outset with OpenAI. It was clear that we would have a core person and point of contact to be able to partner with us and to really listen as well too to what we need to see more from them.

[B]

Nice. So you guys do have a point of contact within them to be able to ask questions and whatnot. Interesting.

[C]

Yes. Yeah.

[A]

What's that person's title?

[C]

I'd have to check. I don't know.

[A]

Do you know are they specifically dedicated to your account or are they handling all advertisers or certain verticals?

[C]

I don't know. That's. I'm not sure how they're managing it on the back end, but I know and at this stage in the pilot, they're really working most with my media team. I don't know beyond just the engagement we've had with them of how they're

[A]

thinking through that and is that weekly calls, weekly meetings. I'm curious what the level of engagement has been.

[C]

I don't have much to share there. I don't know a ton exactly about how it's managing with my team. I'm just getting recurring updates from them.

[B]

Have we should have probably started with this, but we're here now. That's what matters. Have the ads started running yet? Have you had the opportunity to see them in the Wild.

[C]

We have. It was really exciting. They started running. You'll have to look and kind of confirm with when OpenAI released it, but we did start seeing our ads run pretty shortly thereafter and it was just an employee at our team caught one in the wild, which was great to see. They were doing some dog fooding themselves just to engage with ChatGPT about wedding planning generally and it was a pleasant surprise.

[B]

Tell me what they look like. I'm not and Tim's going to wag his finger at me. I'm not as avid as a chatgpt user as chatgpt user as I should be, so. So tell me a little bit about what the ads look like. I know they kind of are appearing below the responses, but what do they look like?

[C]

You can see they have a visual that will do it much better justice than what I will attempt to describe for you right now. But they're fairly stand standard ad unit in terms of like a relatively small placement. For us, we focused our ad campaigns specifically on Vendor Marketplace, which is one of the products that we offer. So if you're not familiar with the knot worldwide, we offer a number of different products and services to help assist couples on their wedding planning journey. One of those is called our Vendor Marketplace where you can be matched effectively or search and find different vendors, local florists, photographers, DJs to help bring your wedding to life. And what we saw, I guess stepping back even further through our own research and working with third party tools that engage with LLMs, is that the number one use case for consumers is searching on AI search to learn about what vendors they should use for their wedding. So it was a pretty natural user behavior, the actual ad unit itself, bringing those two together. So the ads. To get back to your question, Kimiko, the ads that you'll see from us will all be centered around matching you with different vendors.

[A]

And what were the deliverables from your side like? Did you provide copy? Did you provide images? Was it like a single set of copy and images or multiple?

[C]

I can't get into like the nitty gritty of it, of course, but yes, we provided copy and imagery just for not disclosing kind of the full scope of the ad campaign. But we did provide copy and imagery.

[A]

Was that like new copy and new imagery or were you able to pull from existing ads running elsewhere?

[C]

We did. We were able to pull most of the imagery just from, you know, shoots that we've done in the past that were related to similar campaigns. And then the copy we customized for through our in house copywriting team for the campaign.

[B]

And I know that you said that it's only recently launched in this pilot, but have you guys started getting any insights as to how they're performing yet?

[C]

Yes, the team got our first batch of insights just this week.

[B]

Nice.

[A]

What were the insights?

[C]

We're still taking them in and kind of contextualizing them. So I don't have much I can share on the insights just yet, but overall we'll be looking at them through a lens of learning more than anything. Our assumption going into it, and this is not speaking to anything that we're seeing, but it's not that it's going to be like a fully optimized channel out of the gate. We've been, we've been spending on Google for almost a decade now. Google's been highly refined and trained to serve our brand needs. So we'll be looking at it through the lens of like this is early days. We're going to see, we're hopefully going to learn a lot about what's working and then we're going to learn a lot about what's not. And like any good marketer, then you'll optimize from there.

[A]

And you had mentioned like, you know, one of your big questions heading into this was what's the data we're going to be getting back from OpenAI? What are the categories of metrics you're getting?

[C]

So I can't disclose those details just given the confidentiality, the confidential nature of the partnership. So I can't get into things like the data we'll be receiving or deal structure and things like that.

[B]

Would you be able to tell us a little bit about. Because one of the things that interests me here, right. Is that it seems that ChatGPT is kind of propositioned as a brand marketing rather than like a performance marketing play. But how are you guys thinking about it? Is this like a search play? Is it a brand awareness play? How are you thinking about it?

[C]

Such a great question. So I am of the school of marketing thought that brand is performance and performance is brand. So we tend not to get too stuck in the binary of this is brand, this is performance in the way that we think about our media allocations. Why is because ultimately it's all the same to one consumer. A consumer is not looking at an ad and saying like what a great brand marketing campaign or like, oh, there's a performance campaign, I must click on it. So for us, ideally, all the media that we buy is serving both purposes at some point in the journey and Our performance should be working hard for us from a brand standpoint and hopefully vice versa. For this specific one, I would say again, it's too soon to say we know. And I don't say this. We know this through data. I know this through pure intuition that anytime the knot is referenced in a ChatGPT query, I.e. a net positive for our brand. And so of course it is hopefully fulfilling some type of a brand bucket somewhere along the way. It's showcasing our authority, our credibility. Hopefully it's driving some consideration of using our products down the line, if you haven't already. But we would love to see that as they further develop the ad product that ultimately convert as well too. But I think Regardless, the organic ChatGPT queries already, even pre ads, I think we're serving a really important role in

[B]

brand building and then follow up to that for this initiative. Did you guys tap experimental dollars, Brand dollars, Search dollars, Test and learn dollars. Yeah, Tested Learn dollars. Got. Got it. Got it.

[C]

Yeah, we have, we have like a nice, we try and every year carve out a bit of a slush fund that's just reserved for new channels. Interesting media partnerships that may come up. I don't know that we were expecting to spend it so early in the year on something like this. We didn't spend the full budget, but this tapped into those dollars.

[A]

What have you found is working so far with these ads? What have been the big takeaways that I imagine you'll be able to apply to them going forward as well as maybe can apply to where you're running ads elsewhere.

[C]

So too soon to say just because it is such early days. I wish I had more that I could give insight to there. I can perhaps touch on, like what I hope will work for us within this, if that's helpful.

[A]

Yeah, please.

[C]

So I think what I, My, my hope is that we have, we have the largest, the largest marketplace of wedding vendors in the world. We're, we're a global company. We have about 900,000 small businesses that are listed on our marketplace. And finding a vendor for your wedding is a big deal. Like a wedding's a really important moment in time. You want it to be the right match. And I think that connecting like we know there is user demand, we know there's people looking for this and connecting that with this incredibly valuable group of small businesses that we have partnership with will hopefully help just drive more value back to those vendors themselves. And so that ultimately is whenever we're advertising our vendor marketplace is what we're trying to do. And so our hope is that this will allow us to do just that. As I mentioned before, through our third party data, we know there is real appetite for this with consumers. They're trying to figure out who, you know, they want a bohemian wedding in Boston and they want the perfect florist to match that vibe and they're going to ChatGPT to find that answer. And so we're hoping that they will start to see the value and then moving that, that interest over to the Knot itself to be able to browse more directly within that curated library of vendors.

[B]

Now, ChatGPT is not the first to offer ads within its platform. Perplexity had also launched ads, although they have now been paused and then sunset rip but curious if that was an interest for you guys at all. How long have you kind of been thinking about wanting, you know, an ad environment in these LLMs?

[C]

Yeah, you know, we weren't as, we weren't as plugged in on the Perplexity launch. I would say we probably would have been open to it, but we were Most focused on ChatGPT as the biggest player at the time. And I think part of the interest in that particular platform for us was also spurred by the fact that we were already building within there to build the Knot app to be able to call upon within ChatGPT itself. So like our Eng team was engaged, our product team was engaged as we were building that out. And we just know that Surface and platform better. We have been, of course, you know, we've been as I mentioned, a partner of Google's for many, many years and been running SEM through them as part of that. You know, we, we of course are starting to see some of our advertising come up in AI searches as well too. So we are, we also know that it's happening over there, but just in a slightly different format.

[A]

What are you learning there with ads running in AI overviews and like even they're starting to run in AI mode as well. How do those compare with the ads that you're run?

[C]

Traditional search, you know, it's, it's hard to tell because of the way the data that we're getting is structured in terms of, I don't know if you've spoken to other marketers about this yet, but it's, it's, it's not as straightforward as this ran in an AI spot or this ran in an organic search spot. So I, you know, we're obviously, we talk to Google all the time and hopefully as they, as the product continues to develop for them, we'll be able to get some more insight into the difference between the two units.

[A]

Yeah. Kimiko and I had Dan Taylor from Google on the show back in May of last year after they announced testing ads in AI mode and running ads in AI overviews. And that was one of the questions we had for him is what data are you giving to advertisers for them to know when their ads are running in AI overviews and AI mode and they weren't at the time, which is, I imagine, frustrating.

[C]

Yes, I think I get it. I've been, I've been in this, in the marketing world for long enough, in the tech world for long enough to know, like, these things take time. You know, as you roll them out, it's going to keep getting better and better and better. You know, the Google Ad product today looks entirely different than it did 20 years ago. And so I think because of that, I have the perspective of like, we'll go with them on this journey as long as it's something that we think our consumers are seeking. And that's part of, I think, being a voice for the customer.

[B]

How does the going back to the idea of like, data and insights and how much you're able to know, I'm curious, kind of like, how does the traditional search strategy that you have Right. Compare to the one that you're employing for the ChatGPT ads? Is it different?

[C]

A little bit. So I guess I'll stepping back and give you a little bit of insight into how we think about search as part of our marketing strategy overall. And it's honestly been a pretty big transformation in the past year, in part spurred by some of the LLM search dynamics. But in probably January 2025, I actually rebranded our internal SEO team for the acronym to stand for Search Everywhere Optimization and broaden their aperture to think beyond just Google. And at the time, like, we absolutely put LLM search on that as part of that aperture, but also things like social search and p. Because we know users aren't just going to Google anymore to do their searching. And that search like behavior has been a core part of our marketing strategy for the last couple decades. So as the nature of search evolves, we have to evolve with it. That team, who really had kind of built their career in SEO in particular, like, just incredibly fluent in that world, really rose to the occasion to start thinking about those platforms differently and obviously brought in some new tooling so that we were able to get a greater depth of insight. And we started to try and tinker and understand how they work so that we could drive visibility from an organic standpoint. I don't even know if we're calling it organic visibility or not. I think organic search, organic visibility. But the. So I'd say the team honestly has had a lot of fun with it. It's a really strong cross functional intra marketing group that we've brought together that's just been sort of playing in the space to see what works and to drive that visibility up. And so I think we've thought about that. The same people that think about our traditional Google SEO strategy are thinking about how we show up in these AI search engines as well too. But the actual strategies that we're deploying to drive more visibility on those platforms is obviously a little bit different. So while it's the same people with that holistic lens, how we execute those strategies differ by platform.

[A]

Yeah. And so I mean you all then January 2025, we're really early to two of my favorite, least favorite new acronyms, a GEO Answer Engine Optimization, Generative Engine Optimization. But it feels like so many brands as well as publishers and kind of all of us are trying to understand a ogeo basically how to optimize for getting traffic or getting mentions in these various AI chatbots. I was at an event earlier this year where an exec from Mondelez was talking about how Oreo, they saw that Oreo wasn't getting cited that often in ChatGPT queries. And so they were trying to understand, okay, why is this, you know, Oreo is, you know, they said, the number one cookie in the United States. Why would it be, you know, not appearing in these chat conversations as much? They realized, oh, that's because on our site we're blocking all bots and so therefore it just can't get indexed. So everyone is learning like, what are the best practices? How do workflows need to evolve for aog? Given that you all are more than a year into this work, what are the best practices that you and your team have learned?

[C]

Yeah, well, first of all, Oreo is absolutely the best cookie in the United States. I just also a super fan of Oreos. But the. I would say so, yes, I think that the team really was kind of on top of this early in. And part of why, frankly is we have built, we've built a lot of our marketing strategy on search in the past and you know, add traditional advertising and you know, social plays a big role in that strategy as well too and continuing to grow as part of our overall mix. But a lot of people I mean, think about the user behavior. Again, it all comes back to the customer. You get engaged, where do you go first? A lot of people turn to things like Google to be like, how do I start wedding planning? Like, where do I go? We prefer they come to us first, but we understand that that's part of the user journey at times. So with that in mind, I think the good news is we were pretty well set up for the advent of AEO geo. And why is because we have built such credibility and authority on Google over such a long period of time. So we already were indexing quite a bit of our content to ensure that it was surfacing in those environments. And so as the LLM engine started feeding that in, it was able to read through that content fairly well. So our visibility out the gate was reasonably strong. How we have. So one is just, it is exactly as that executive said, you know, like ensuring, like doing some of the dirty work in understanding, like what are you allowing them to crawl? What do you not want them to crawl to is an important component in this. So if you need gating or bot blocking set up, making sure that you're being protective of the data that you want to ensure that is truly proprietary as well too. So one is ensuring that you have that smart crawler bot strategy enabled so that you can index and at least have a shot on goal. The second, I would say is trying to really understand the user behavior. And if you think about it, the difference with AEO is like you're answering questions, you're not just surfacing laundry lists of data. And questions are more specific in nature. You might not answer it with an entire big article. So I think taking a consumer first lens of what are people asking and really trying to understand that we did well partner with a third party tool. Have you heard of profound? Has that come up? Oh yeah, yeah. They seem to be kind of taking the lead in the industry. And so we levered Profound data and have found it to be tremendously helpful in giving us insight into different categories that people are searching within. Of course you don't get down to the question level, but you get broad strokes. So then from there you're able to then kind of scrub through your content of like, how are we answering questions that would naturally emerge within all these different categories that are spiking for usage? And then for us, of course we have a massive content library. We've been in the content and editorial world for a really long time, so it allowed for us to look at that content library with some fresh Eyes of like, are these answering the questions that are most likely coming up and making sure we're refreshing content with this, this new platform in mind? The third, I would say, is like thinking beyond your own content as well too. And what is your footprint across other parts of the Internet that might be informing those decisions? And so forums like Reddit, press releases, articles, all those things are funneling in and helping drive informed decisions through the algorithms. And so that we're probably a little bit less focused on at the outset. We're just kind of continuing to pursue our press strategy as we do and make sure. We're being really mindful of making sure that we're putting out all the good work and innovation that's happening within the company. But I'd say that's another area that I would say as folks are starting to look at this more seriously to look at as well too. Even things like trustpilot, how that could potentially be influencing how you're indexing, particularly from a sentiment standpoint, merits at least a hard look at and potentially a whole strategy against thinking about how you would want to be showing up in those environments.

[A]

Yeah, it's like AI turns everything that was, you know, two dimensional or three dimensional into like 5D chess at this point, because you're having to think about, to your point, content marketing in a different way than I would think you would historically.

[C]

Totally. It's like, you know, they used to say, like, Google's the front door and now it's like you have a front door, a side door, a garage door, a back door.

[A]

You don't have walls.

[C]

Yeah, no walls. You're just in like the middle of a campsite, which is. I know. So it's just how you figure out what are those the biggest doors and how are you influencing those. The other thing too, that I should have mentioned on your own to proprietary content is something that we've been thinking a lot about is Help center as well too. Like kind of the underdog of the content world, but man, such a valuable piece of real estate, especially when it comes to answering questions. And hopefully our users are able to get what they need from our Help center directly. But in the event that they're turning to ChatGPT or something like that, ensuring that our Help center is well optimized to be able to answer those questions in that environment too.

[A]

Have you all made any adjustments to the article schema of the Help Center? I feel like FAQ has become the popular format for AI bot crawlers, but then also things being written in Markdown like Cloudflare. I think Cloudflare is the one who's popularizing like Markdown for AI agents now. XML is like back in vogue again, which is just the ugliest text format, but very not ugly for an AI bot.

[C]

Yeah, you know, I honestly am not entirely sure. I'd have to check on that one. If we've had to make changes. Some of that work on schemas and bots and so forth happened mostly over the summer. And admittedly I was on mat leave over the summer, so I missed a little bit of like the, the nitty gritty on that. But. But if that's of interest, we could, we could certainly circle back and let you know.

[B]

I'd be curious on the other side what this kind of means for dollars. Right? Because going back to your point about it becoming like a content marketing engine and things like this, and even extending the social, like does the search buzz, it just get bigger to encompass more things here, or are the divvied up or how did you guys adjust the budget in light of these changes?

[C]

You think our advertising budget or content? I'd say I don't know that we've adjusted it massively in light of these changes. The media mission that we've been on the past couple years has been mostly around how to think about just again, it sounds back to basics, but being where our couples are. And so social has been a big part of our media strategy over the past few years. Partnering with creators leaning harder into platforms like TikTok, like Pinterest, like Instagram, to ensure that like, as they're searching in those environments as well too, we're showing up. So I'd say as I look at our media budget and allocations overall, we've seen a bit more of a skew towards social than we have in the past. Like, that has been an intentional part of our strategy to ensure that we're well diversified across channels. But I wouldn't say that was spurred by AI search necessarily more by consumer behavior at large.

[B]

Got it, got it, got it. Well, I guess we'll continue. I hate to say watch this space,

[C]

but, but watch it, watch it. Kamiko.

[A]

That said, I'm curious. So like you mentioned, being where the consumers are with search, you can only be there to like a certain point, basically have the domain or the ad to then get people to come to your property or to your app. With something like ChatGPT, there's newer opportunities to be even more present there. We've talked about ads, but also through chatgpt apps which the Knot rolled out a ChatGPT app earlier this month, earlier in February, which still it seems like that's a traffic driver opportunity. But because I think if I'm not mistaken, from playing around with it, the idea is I can invoke at the knot within a ChatGPT conversation and then that'll pull from the Knot for me to be able to find a vendor. But then I would click from there to go to the Knots website or web app for it. But what was the thinking here? Because similarly with the ads, it's like ChatGPT apps are very early. I think it was only in October of last year that OpenAI even announced they were going to have these as an opportunity. So opportunity to get in there early. At the same time, consumers may not even be aware that this exists. So what's been the marketing around the ChatGPT app that the Knot has?

[C]

Yeah, we haven't so. Well first, thank you for playing around with it. I'm glad. I'll go look at our numbers and hopefully see a little uptick in San Diego. I'd say the marketing around it has been pretty light for the app itself and that's partially just because it's new and like any good beta product, you're not going to do a massive go to market to promote it. We just launched ours at the end of January, beginning of February, so it's early days. But part of why we prioritized it is again just to allow for a more native search experience. Weddings are a highly visual space, allowing for some of the visuals from our vendors storefronts to actually live more natively within. ChatGPT itself just allows for couples to make more informed decisions about finding wedding vendors that feel like the right fit for their esthetic. So from but in terms of marketing to allow people to your point to call upon the Knot app, we did some some promotion on social. We've I believe we ended up sending some form of a mention in some email, but it's been pretty light overall. We're not putting paid dollars behind it yet. We're mostly just trying to learn and see how users are going to take to it. And as we continue to tinker with making it work better and better and better on the back end, what are

[B]

the learnings that you guys are hoping to kind of get from this? I think the reason that I asked is because AI moves really, really quickly in terms of like features, availability, how people use it. So given that it is early days kind of what are some of the expectations around this app,

[C]

around the app. I would say I think what we're really trying to most understand is user adoption and behavior, especially with such light marketing of just how organically are people going to that app marketplace within ChatGPT, using the Knot app, calling upon it and trying to see more of the organic and evergreen adoption of it versus us, certainly before we put a bunch of paid dollars behind it, but without having to put a bunch of like just really trying to understand user behavior and then how does that ultimately convert back to site, whether that be an app download, whether that be going to our vendor marketplace and submitting a lead to a vendor there. But of course, you know, one of our end goals is to drive value back to our products. So I think we'll see as that continues to come to light of just how that user behavior evolves. But we want to be on the front lines of where innovation is happening. And clearly ChatGPT is one of those places.

[A]

You all are definitely on the front lines. One of the first brands to advertise on ChatGPT, a brand that was thinking about AEO and GEO in January of 2025 when you all started with that. So, Jenny, this is really, really fascinating conversation. Thanks so much for joining Kamiko and I.

[C]

That's awesome. Thanks so much. This has been really fun. I appreciate all the great the interest in this space. You know, we're having a lot of fun with it and I just find it to be like a really fun intellectual challenge and this, this new advertising era. So appreciate you having me on and loved, loved the talk.

[A]

Thanks for listening to this episode of the Digiday Podcast. If you enjoyed it, please leave us a rating and a review on Apple Podcasts, Spotify or wherever you're listening.

[B]

Listening.

[A]

Get more from Digiday with our daily newsletter sent out each weekday morning. Visit digiday.comnewsletters to sign up.