Kelly Wortham: All right, welcome back, everybody, to conversation number 110. We have another panel, and it's a crossover with No Hacks. Yay, welcome back, Sani! And this time, we're talking brand trust, AI, and what's breaking. We've got Juliana Jackson and Wil Reynolds here.

Kelly Wortham: But before we do our introductions, I mean, not that they need introductions, but we always have to do that, because we want to make sure anybody who's new to the crew knows who we're talking to. But first, we want to thank our sponsors, as always. Let's start by thanking Conductrics. They've been with us from the very beginning. Thank you, Matt, Nate, Ezekiel, Ashley, and Dominica, and the entire team at Conductrics so much for supporting the TLC.

Kelly Wortham: We also want to thank our other TLC newsletter sponsors, starting again with Conductrics and GrowthBook and Ask Y. Thank you all so much for providing your thought leadership to the community. The learnings we get from you makes all of us just a little bit smarter, so thank you so much. Thank you, Convert, for sponsoring the amazing ConvertStations channel. It helps all of us get to know each other. If you haven't gone over to the Conversations channel to answer some of those questions, I really encourage you to do so. It's lovely to read your answers each month, throughout the month, and taking a chance to answer those questions gives you a chance to maybe win a little bit of money. And last, but not least, thank you, Eddie Aguilar and Blazing Growth.

Kelly Wortham: So, who's here? Make sure I got them all right this time. I did. Okay, so we've got my co-moderator, and my good friend, Sani. He's the host of No Hacks, and one of my favorite people to have real conversations with about where AI experimentation and digital experiences are heading. He has a way of cutting through all that noise and getting to what actually matters, so welcome back, Sani.

Kelly Wortham: Juliana Jackson, we just had her here, but, you know, you can never have too much Juliana. Like I said last time, if you're not reading her newsletter, Beyond the Mean, you are missing out, so go to julianajackson.substack.com. She's one of the sharpest voices I know when it comes to all this stuff, but especially AI, customer understanding, what brands are missing right now. She brings the perfect mix of strategic thinking, honesty, and that's-not-how-that-works energy! So definitely go check out Juliana's work.

Kelly Wortham: And then Wil. Wil Reynolds, he is the founder of Seer Interactive. He has been challenging the industry to rethink what visibility, search, and trust mean in the AI era, along with all of his team. His perspective on AI discovery, authenticity, and the future of the web is one of the most important conversations happening right now.

Kelly Wortham: So, to get started, I think I'm gonna throw this to you, Wil, first, and then I want everybody to answer. So Wil, for everybody listening that maybe don't even realize that there's a problem yet, can you kick us off by explaining what the shift is that's happening that should be setting off alarm bells?

Wil Reynolds: Well, jeez. Well, first of all, thanks for having me. I think the shift that I want to talk about is, because I think now people get that, like, okay, people are using AI to find stuff. You know, it's like, yes, they are. I think the biggest shift that I'm seeing is that people are focused on prompt tracking and AI optimization in ways that are not long-term sustainable, and I think are putting their brands at risk. And we're already starting to see early signs of penalization for some of these things, and people really need to start to ask themselves, was it worth it to get those results a month or two faster for the potential to have your brand or your domain penalized in at least Google for now, because Google knows how to penalize companies for doing low-quality stuff? It's only going to be a matter of time before they layer that into Gemini and other places. So that's the thing, I think it's like, wake up, do not let these consultants help you with your shit, and then tank your brand someday, because they get to slide off like Homer Simpson in the bushes, while you're stuck dealing with a brand that now has a big X mark on it across the major platforms that people are using. So I would be very cautious about that.

Kelly Wortham: So you're talking here about essentially doing the same dark pattern, or similar dark pattern tricks that they would use in the SEO world, to make your site appear in AI prompts.

Wil Reynolds: 100%. Like, there are, I mean, the white text, white background stuff, I think people are pretty much like, we probably shouldn't do that. But the, let's launch 50 pages on one day that follow the exact same pattern, people are like, yeah, that's great. And I think one of the things that we have to rectify is the speed versus the risk. If you're gonna sell your business at a certain point in the next 6 months, and you're gonna let that be someone else's problem, go ahead and do all the tricks. But for a lot of the companies and organizations we work with, they need to be here longer than that, and they need to be visible in these places more than 6 months. So we have to have the chutzpah to be able to kind of tell people, like, yes, I could do these things, yes, I know how they work, but I don't believe it's right for your brand. Come back to me in 6 months when you're one of the hundreds of people that got hit as a result. Already seeing it.

Slobodan Manić: Mount AI.

Wil Reynolds: Mount AI. Go look at airbyte.com.

Slobodan Manić: You can look at their bite.

Wil Reynolds: Back on.

Slobodan Manić: It's still a case study on the website, right? A successful case study.

Wil Reynolds: This is… so, why are you making me mad this early in the morning? So, type in AirByte Case Study. Air BYTE case study, you will find two AI optimization agencies who have them up as case studies, then take AirByte and put it into SEMrush, or put it into Ahrefs, and look at their pattern of pages indexed, and they do this. How in the world are these agencies working at such a fast scale that they're not realizing that the case studies they're putting publicly on the web have clients who have been penalized by their exact same tactics? And they're just letting that sit. And how are we as buyers being that stupid that we're not taking the case studies we're seeing, and then putting them into these tools to be like, wait a second, they just had a massive decline, where they lost 80% of their pages indexed in Google. Are you sure that's a case study? And what does that say about you as a business, that you're not even checking if your work is helping clients just a month or two after you did the work? You're not a partner, you're just doing the land grab.

Slobodan Manić: We're 5 minutes in. I just want to say that I love this already. But, yeah, so very true.

Kelly Wortham: So, so what is the risk, and what do we need to change about measurement? So what do you need… how do you do it right?

Wil Reynolds: Oh my god, so this is the sad part. Measurement requires education, bravery, and honesty. A lot of our leaders are saying, hey, I read this case study, it said this. Great, let me show you what they did. Do we want to do these things? Yes, no. It's education, right? It's like, you know, we can say to our… to the cows come home, like, oh, well, every time you prompt, it gives a different answer. That's not a good answer to a C-level or a board member, right? Like, the pressure's coming down to be like, I need to show up, I searched for this thing, why don't I show up? We have to be able to take that pressure and lead instead of follow, and say, well, let me show you why you might or might not be showing up. Let me show you how we can solve for that. Let me show you what's roadblocking me from that. If you unleash these roadblocks, maybe I can help to fix this, right, in a sustainable way for the long term. I think so many of us are getting pressure also for ROI. ROI is horrible. Like, I looked at all of our calls, like, over probably 100 calls, where people were saying they want to show up higher in AI search. The amount of them that also then said, and I need to know the ROI, was, like, 75%. So, you want to invest in an experimental channel that didn't exist just X long ago, and there's no monthly search volume, so we don't know what people's prompts look like. And the prompts are personalized, so we have no idea if the answers are how consistent they are. And Google does dumb stuff, like, they made all their Gemini traffic direct a couple of weeks ago, and now all that's not coming in right. And you're tracking the wrong prompts, because you haven't… you made up prompts instead of actually using your call transcripts to figure out what real customers are saying. So you start stacking all this on top of it, and you go, you want me to give you an ROI? And there's no clicks! And a lot of times, brands are mentioned in so many different ways that sometimes you can click on their answer and go to their website. Sometimes they just list the brand, which means you have to Google it, which means it comes in as organic direct. It's like, it is so… there are so many caveats to how do we analyze this stuff, that if we don't take the time to educate our leaders on how to track it, and what is and isn't possible anymore, we're allowing them to live in a world of, wasn't all this trackable with SEO? And we're not pushing back against that. So I think that's really, really important.

Slobodan Manić: I agree. Juliana, you're the editor in my life. You're calling out bullshit wherever you see bullshit. What do you think about all this?

Juliana Jackson: I mean, I agree with Wil in from a tactical perspective, this is what's happening, and I'm working in an agency too, so we're very likely to see the same stories. I will say that from a macro perspective, what we… I guess the shift is that all of the big tech platforms, like Google, like Meta, they're just retraining us on how to search and engage on the internet. That's kind of, like, what's going on. Everybody right now with the phone is being retrained into how they find information, how curiosity happens, and I think the stakes right now for… Wait, let me reformulate this, because I'm trying to stay employed. Okay, yes, so the stakes are high for tech giants, which I will not name for this call. I think these tech giants are the ones that are retraining us in the way we should search, in the way we should be curious, and in the way we're going to, I guess, from that participation to intent to cart, and so on. I think there's a lot of top-down AI mandates that are happening in companies, which makes people stressed, which there's a lot of pressure on people to attach AI to every little shit that they have. Does it help? Probably not. Most likely not. So I guess it's just a very… it's like an inertia that I'm noticing as a person that works in measurement. I have a, you know, my experience is in measurement. Of course, I cringe, I'm sad, I'm, you know, I go and I write things up, but I am still excited more than ever for search, because search now is basically the experience. Like, forget about websites. I see the search as the experience, like, how do you get people from curiosity to actually take an action. For me, it's very interesting, because I recently, in the last few months, I've been working mostly with YouTube and creative, and I use the creative as the attraction, as the targeting, as the opening door. I'm not that much into SEO, but I do understand the value of brand perception, I do understand the value of share of model, and so on. But I'm mostly focused on the creative assets, and how can you create really good… and I heard this at GML yesterday, the best ads should be answers, so it kind of shows, in my opinion, how much search influences the way we communicate and our messaging, so I really related to what they said at GML. The best ads are answers, which shows the influence of search, which shows the influence of how we're changing our habits around how we are asking questions, how we are finding things. So that's where I'm still being positive so far, but that's kind of, like, my TLDR of the shift.

Slobodan Manić: I like that.

Kelly Wortham: Did either one of you, or any of you see Rand Fishkin's post yesterday about what if the future of AI is bifurcated? He talked about consumers say they hate AI, they rate it negatively, adoption rates have slowed. But businesses are going crazy for AI.

Juliana Jackson: It's provocative.

Kelly Wortham: It is! It's really provocative! And I thought the analysis was really interesting. Juliana, you were just pointing out that, like, you're getting this top-down pressure to use AI, use AI, use AI, but if consumers are potentially using it less, what does that mean?

Juliana Jackson: Like, if I'm a numbers person, I would say, what proportion of the population is it using? What is their background? I'm very careful every time when I get numbers, because you can… if you research right now, you will find very interesting things, like 5% of people use AI mode, 45%, then 60%, then 50%, but who are these people?

Kelly Wortham: And what does using it less mean?

Juliana Jackson: What is the sample size? What is their context? For sure, people in middle America that are old and, you know, been doing things, they're maybe not asking recipes from Gemini and figure out how to do a casserole, but come on, like, people that are, you know, working in the industry, or outside of the industry, that have a different lifestyle, so I think it's all very contextual correlation, it's not causation. I know Jim Janoglio's gonna like that with his MMM. But who are these people? Like, Seinfeld says, who are these people? You know?

Slobodan Manić: Yeah.

Juliana Jackson: That's my take. Yeah.

Wil Reynolds: I don't know. Same thing happened early in search. I remember sitting in boardrooms when people were like, so you think that someday people are gonna put their credit cards in the internet. And you're like, I think… yes. And they're like, you're freaking nuts. Yeah. And you're like, here's my report from Gartner, here's my report from somebody that says that people don't feel it's safe. And then, you know, it's like, okay, like, we'll see. A lot of these things are about timing. You know, too many people make predictions without expiration dates. So then they can always say they were fucking right. And if you think about it, so I worked with Peapod back in the late 90s. And they were doing grocery delivery. And they failed and failed, but the people who predicted that grocery delivery was going to be huge back in 1997, 1998, when they raised their billions in funding, now get to be like, see, I was right. It's like, yeah, but you were off by 26 years, dawg. Like, the company went out of business, so I think these predictions and these studies, it's like, who are these people is the right question to ask, and I think we ask it at the highest level, but we don't always want to get to the most micro level. So my version of who are these people sounds like this. When somebody fills out a lead form and says that they use ChatGPT, I email them or send a video to them as soon as I possibly can and say, can you share with me your prompts? Sample size of 1. But at least I get to get a little bit more of an understanding that's not some bullshit in a study. I'm like, for this one person, this is their reality. And this is how they're searching, and this is how they're doing things, and I think that that's just… there's this human component that is not… that I think we're missing, in terms of, like, if you're not watching real people do prompts, you are a victim of what somebody else's study said, with no other context, to be like, that could be right for whoever you… who are these people? Whatever the group of people that you found for your thing was this. People are using it in such different ways, it may not apply to all of us, and you want that, like, grounding in real customer behavior to know, like, that's a great stat, but you know what? My customers seem to be different. Okay.

Juliana Jackson: Perfect.

Slobodan Manić: Small sample.

Juliana Jackson: That's all I have to add to the effect.

Slobodan Manić: But I think the problem is that this was just thrown at people without a manual. Like, everyone has to figure out how to use it. We don't have one way of using any… like, how many people have used AI, LLMs, beyond just the chatbots, global population. It has to be less than 1%. Like, it has to be less than 1% of all the people in the world. So we're talking about 1% of people and arguing about 0.6 versus 0.3% of all the people in the world. Still meaningful, still a sign of things to come, I'm sure, and all that, but we're still talking about an absolute minority, and we're pretending like that's, like, the entire universe. I think that that's important, and also the fact that everyone uses these things in a different way, because it's, like, a very private thing, the way you use AI. Unless a company worker has policies, like, you have to use it this way, that way, this is the workflow. Everyone figures it out on their own, and they just show up and look, and oh, that's not how you do it. It's supposed to be done. If you say, don't lie to me, it's still going to lie to you, and stuff like that. So, I don't know, we need more transparency about how everyone uses these things. That would help everyone.

Kelly Wortham: How close are we to being able to tease out AI-referred versus non-referred traffic? Because you say it's small, but you're also the one who shared the article. Sonny, you're the one, the Adobe data.

Slobodan Manić: Well, GA4 has that now. Like, they have AI referred traffic since, like, last week? Last week, yeah. So, it's showing up, I don't know what the percentage of that traffic is compared to overall traffic, but I'm talking about people in Third World countries, people in, like… Right. They will not be doing this. It's growing, whatever it is, it's growing, and we have to treat it like that.

Wil Reynolds: And it's interesting, the other countries thing, I thought I was hearing that, like, the Americans are the most skeptical ones, and that, like, people in other countries seem to be less skeptical. I don't know where I read that. Did anybody else read that? Does that ring a bell to anybody?

Slobodan Manić: There was a China versus US, and yeah, the skepticism was an American thing, and China trusts AI, like, 70% of people trust AI. I saw that yesterday, somewhere. It's not world, it's China, specifically. Who knows how they source the data? Who knows if that's true? We don't know that. Yeah.

Wil Reynolds: Who are these people?

Slobodan Manić: Exactly, exactly.

Wil Reynolds: That's the title.

Slobodan Manić: But it's, yeah, like you said, Wil, show me your prompts. Let's just start. What do you even do with this thing? And I'll show you, I'll tell you how you need to adapt. But yeah.

Juliana Jackson: I want people to see my prompts, because I asked the AI the most dumb shit ever.

Slobodan Manić: Everyone does, everyone does.

Juliana Jackson: Bro, you don't get it.

Slobodan Manić: No, no, no.

Kelly Wortham: Excuse me, Sani, my prompts are very intelligent.

Slobodan Manić: Hey, most of them are, I'm sure.

Juliana Jackson: Let's move on. I'm gonna be like that guy from DMP.

Wil Reynolds: Right, got it.

Slobodan Manić: A lot of people. You know.

Wil Reynolds: I think one of the things that, when it comes to tracking, so you're asking, Kelly, about referrals and tracking and things like that. If you're not tracking layouts, then you can't explain the changes you see in your analytics platforms. For instance, when there are certain prompts that I do where, like, half of the results for companies to hire or whatever are links to their website. And half of the results in the same answer are just brand name mentions. Like, how… I can't physically click on your brand name, so if your reporting is set up to say, well, I'm winning in AI when my traffic does this, it's like, well, how many of your… how often are they changing the answers, and how they look, and what's linked, and if you don't have that in a database where you're saying, maybe every day I'm taking a thousand random prompts, screenshotting what it looks like, but then you don't have the upwind intelligence to know, oh, on this day they changed the layout, and that's why my things look the way they look, and that's why my traffic is acting the way it's acting. You're just thinking, oh, I did a great job, I guess, because our traffic jumped, or oh my god, I'm in trouble because our traffic is down, and it's like, but you don't even know what happened that led to that, and you're not tracking that anywhere. You know, Google's making a lot of changes to layouts right now, that, you know, you gotta kind of track those on a regular basis. Like, get a screenshot API, and just screenshot it, vibe code it, and just be like, dump it into a folder every day. So when I have these questions, I can now at least say, have you seen a change in the layout that could answer this problem that I'm seeing?

Slobodan Manić: That's great advice. That's really…

Juliana Jackson: They announced at GML that they're gonna add more formats, especially with search and how ads are gonna appear in search, and it's just, like, I'm still unpacking all the shit they announced this week. It's so much going on, like…

Slobodan Manić: Yeah, generative UI is another term that they throw around, like, they'll be doing generative UI in your SERP, and then, like…

Juliana Jackson: Hygienic proof.

Slobodan Manić: Kind of changes everything, yeah. But Wil, the thing, one thing that you talked about, I talked to Elisa about this, what do you know about my brand, or what was the prompt that you were using to figure this out? I think that's a good starting point for everyone. Forget about best X for Y, like, just let's talk about, tell me about brand.

Slobodan Manić: And…

Wil Reynolds: People are addicted, man. It's like, they're addicted to non-branded prompts, man. Like, it's really sad, because what you do is you watch a company spend millions to become a brand and a household name.

Kelly Wortham: On the non-brand stuff.

Wil Reynolds: And then when it comes to AI, they literally are like, where do I show up for best checking account? And one of the examples that I like to use is, when a person was typing in, what's the overdraft fee for Bank of America, and then what's the overdraft fee for Wells Fargo. Wells Fargo owned 7 of the 9 answers that were linked out. Bank of America owns zero. So, all your stadium naming, because what's the thing that people care about most with checking accounts? You know, I don't know what they are, I don't purport to know what they are, but I think one of the things might be, what are the fees like? Making that up. Okay? So, if people care about fees in their accounts, then is what is said about the fees for your brand accurate. And instead, what's happening is, they're just telling me, hey, my CEO's mad, when you type in best checking account, we don't show up. And I go, well, what about the millions you spend for people to know your brand and put your brand in their prompt? You don't have to worry about citations and… Dude, go look at those and make sure they're right, because when you update your rates, you are at the mercy of somebody else updating their website because your site isn't crawlable appropriately to answer the question for the person who already knows your brand. You have spent all this money for them to not type in checking account, but for them to type in Bank of America. You know how much money you've gotta spend to get someone to change the way they search, instead of being generic, being brand-specific? And you're not gonna own that? Own that first. Like, go figure that out first.

Kelly Wortham: Own your brand.

Wil Reynolds: They already know you. They're comparing you to somebody else. Why try to be compared to 50,000 other banks out there when you could compare them to just the other person they're comparing you to? How are we different or better than this competitor? And I know we don't want to write that content, I get that, but is the answer true? Does it put you in the right light? Where is it getting its sources? What can you change? What can't you change? Those things are important.

Slobodan Manić: So what… let's just wrap this trust discussion up. What's the one thing people should be doing, brands should be doing? Start with that prompt, like, what do you know about this brand? What does this brand do? Like, is that a good starting point, or is this something else? If you're not doing anything right now.

Wil Reynolds: I always tell people, just type in, tell me about my brand, and then see if the answers actually match what you believe about your brand.

Slobodan Manić: That's awesome. I like that. And what if they don't own that narrative?

Wil Reynolds: Then you need to take a good look in the mirror and ask yourself, why is it that the things that we believe about our brand are not on our freaking website? That's actually the problem. Like, so, the best example I can give is, one of my alumni hired us at a company, decided to start treating people like shit that were her old co-workers, and we had to fire them as a client. Well, she then, back in 2012, went to every freaking website that she could to review an agency and reviewed us negatively. So you can see the pattern because they all showed up on the exact same day, right? So, okay, you fired me, what is my retribution? The only thing I can do is go to Agency Spy and a bunch of other websites that people don't visit, and slander your brand, not be transparent about what happened, right? This CEO sucks, can you believe the CEO? I'm like, wait, you didn't tell them that you were an alumni that hired me and worked here for 6 years? Like, you're gonna exclude that? But anyway. My point is, is that I could always, in a search world, be like, fuck that. In an AI world, it's looking for pros and cons, so it's trying to find balance, so it goes super deep when it doesn't find something negative, so it started now indexing all that freaking deep content and saying, here's a negative about SEER. I go, okay, it might be true. Maybe we do have crazy account manager turnover. So the first thing is you've got to be honest enough with yourself to be like, maybe the negatives are true. Step one, maybe the negatives about your brand could be true, even if you don't want them to be. So the first question was, oh, how high is my turnover? And then the second question is, well, what's the average rate of turnover for an agency? And I just did deep research, I'm not gonna waste too much of my time doing it, but I did deep research, query, find me some stuff, and I said, alright, we're actually lower than the average. But guess what we never did? We never published our account turnover. So therefore, we allowed these third-party sites to dictate a narrative when people are typing in my brand. Hey, tell me about Seer Interactive. Oh, you know, they're great for all these things. Their account manager turnover is really high. I'm like, well, damn, it's all coming from this one set of sources that were all done on the same day. The AI's not smart enough to know I've been in existence for 24 years. If you only got one negative review, that's a good thing. It's not that smart, right? So you see all these things stack up, and I'm like, well, why don't we just publish our turnover publicly, along with the research that we did to show, like, hey, this is the query we used in deep research. Here's what the turnover is. We changed all the answers about our brand. And it turned it into a positive. But if I wasn't willing to, one, just type in my brand, two, look at the answer and see, like, is that true or not, three, then publish it publicly. So now it's… now I've got a… now I'm in their arena. Hey, I got content in the world now that talks about my account turnover, and now my responsibility is just to keep it up to date month over month, year over year, so that it's not some old shit when it looked good, and if it gets worse, now I'm like, well, we're just gonna leave up the old data. Like, you can't be like that.

Slobodan Manić: No, absolutely. Juliana, you're nodding the whole time.

Juliana Jackson: Yeah, I mean, I'm in this webinar to open for Wil, so I'm, like, the opening act, so I'm just listening. I'm a big fan of him, so I'm just like, here, you know, like, damn, that's true. I'm not a search person, right? Like, I'm…

Slobodan Manić: This is not search, this is beyond search now.

Juliana Jackson: It's a lot of… when I was listening to you, Wil, like, all I keep on thinking, when I was at Monks a few years ago, this whole… yeah, you know, this is gonna be… no, no worries, no trashing today. I'm gonna keep the kosher. But, listen, I was at Monks, and this was before, obviously, Jellyfish and Share of Model and so on, and we were talking about how do you appear in this model, so it was AI overviews back… not even AI overviews, it was just shit was just getting started. So I remember they were trying to engineer all these responses in the LLMs, I said, but why don't we just… why don't we just do marketing? Like, this is… we should do more PR, we should do more content, we should do more… like, we should all try to do, you know… and they were like, no, this is about prompt engineering and response, and I'm like, I don't think that's what… I don't know. Then, when I got to Jellyfish, and we have Share of Model, and I'm not… I don't work in that area at all, but they have been doing a lot of good stuff. It's all about what Wil is saying, it's about brand perception, it's about coming to terms with how your brand…

Kelly Wortham: And authenticity.

Juliana Jackson: Is trusted. So, I mean, for me, if you have… you're not doing anything today, just do marketing, like, that's just do decent… I know it's provocative, it's crazy, but write some quality articles, do some interviews. I don't know, publish in different newsletters. Publish your newsletter that you send, publish it online. Get more stuff. Find some random people to do exchange and backlinks, like you used to do it, like it's 2015, I don't know.

Kelly Wortham: I love that you say, I'm not in search. Like, I've just started writing about all of this stuff, because smart people like you all are teaching me as I'm trying to learn. And I literally had somebody on the Measure Slack, like, ping me and send a link to my article to me and say, this is really great content, why haven't you indexed it? And I'm like…

Slobodan Manić: Oh my god.

Kelly Wortham: What does that mean? Not in search! And then…

Juliana Jackson: So they had to be a league.

Kelly Wortham: So that I can start indexing my own articles.

Juliana Jackson: Listen, if you put Search Console on your Substack, you're entering in the world of hurt, because, you know, it's…

Slobodan Manić: Dark. Let's not go there, let's not go there.

Juliana Jackson: Oh, no. Let me talk about it. I have time today.

Slobodan Manić: Let's spend 2 minutes on it, let's spend 2 minutes on it. I have a question for Wil.

Juliana Jackson: So I decided a few months ago, I mean, actually the last year, that I'm gonna add Search Console to my Substack, because I was feeling myself. So when I do that, I start seeing full-on prompts from ChatGPT, indexed. I was seeing people's email address, their resumes, their planning, their mortgage, like, there was a whole thing.

Kelly Wortham: Oh…

Slobodan Manić: Cheating, boyfriends, girlfriends, stuff, all kinds of stuff, yeah.

Juliana Jackson: And I said to myself, listen, like, I didn't ask for this data, I don't want to be collecting this data, so I deleted my Search Console, so me and Search Console are… because I couldn't take it anymore. Like, it was people's name, I looked them up on LinkedIn. I was like, holy… this is crazy, like, whole addresses, and people really use, like, there's so much anthropomorphization. I said the word right, poof. It's so much anthropomorphization with these models, and, like, I don't want to be part of that at all, so that's why I'm telling you, talking about indexing, Substack will index you automatically, because it's their… because you're just part of their domain. You just need people to refer you and name you and, like, oh, shut up.

Kelly Wortham: Not on Sunday.

Slobodan Manić: Oh, but that was a crazy story, that was just a crazy story.

Juliana Jackson: I've seen things, I've seen things.

Kelly Wortham: You cannot unsee.

Wil Reynolds: Seen things.

Slobodan Manić: No, it was… it was everywhere. The story was covered by Ars Technica, by all the…

Kelly Wortham: It went viral.

Slobodan Manić: It was everywhere.

Juliana Jackson: Legitimate.

Slobodan Manić: 100 plus countries wrote about it. Yo, I tried…

Wil Reynolds: When I came… when I… I was so angry, because when someone mentioned it internally at SEER, I was traveling. And by the time I got to where I was… I got to, because my brain was, like, cooking, and I'm like, we should run site colon searches on ChatGPT's shared answers, and scrape all that with Screaming Frog instantly, because then I would have had a database of how people prompt, so I can better understand people. And I landed, and they turned all that shit off, and by the time I went to go fire it up, it was, like, a day later, and it was like, couldn't do it.

Slobodan Manić: That was August, this was October. This is a different leak. This is even worse. This is even worse, yeah.

Juliana Jackson: LLM prompts in my Search Console.

Wil Reynolds: Oh, in your Search Console!

Slobodan Manić: Yeah, because they use Google to search.

Wil Reynolds: Juliana, wait, you're the one that fucking leaked that?

Kelly Wortham: No, allegedly.

Juliana Jackson: I was not the one. It was Slobodan and Jason Packer.

Slobodan Manić: That was us. That was us.

Wil Reynolds: I remember that, because I was like, it's in Google Search Console, but I didn't realize where it came from.

Slobodan Manić: That was us.

Juliana Jackson: I was done, yes.

Slobodan Manić: Go back to when we did the interview a few months ago, you told me about trusting an expert in an industry. You mentioned the GEO bros, like, talking about how they know all the shit, and just the example you used is, I go to every single conference in the industry, I never see you there. Like, they would invite you to speak. Like, it's easy to fake expertise online, but AI kind of exposes the fake experts. Is that a good thing? You say that, good.

Wil Reynolds: So let me share this with you. I can't believe I actually have it open right now. So, Hacker Noon, right? Everybody here know Hacker Noon? Alright. It's, like, a decently reputable site, so I thought. So I'm gonna put a link in the chat. And, so I got an alert the other day that's like, oh, you got a mention on this list, and I'm like, okay. And then I see at number 10, this guy Evan Balen, who I've just been railing on, because he does a bunch of listicle, low-quality shit where he puts himself as the top GEO person. So I'm thinking this domain and this brand, branding, marketing, regular, everyday marketing. You've done good stuff, Hacker Noon. There's some stuff on your site that I have loved, so therefore I've always held you in a certain esteem. Then you allowed this to be posted, and the number 10 person, Sani, is exactly what you said. It's a person that has spoken at zero of the conferences I've ever been at. But yet, they're ranked number 10, and then the person who wrote it put herself at, like, number 4, and you're like, this shit ended up on the site? Like, if AI wasn't a thing, would you have written this? And it's like, no. Right? Like, there's this disconnect where people in the industry would… or even buyers, people buying services would be like, are you just blindly looking at these lists and being like, well, this person's a top person, so I'm just gonna start following them? Go ahead and follow Evan Balen, see what you learn, other than he's awesome, right? And it's like, he's not putting anything out there, he's not moving the industry forward, there's nothing you've ever shared from this guy to anybody you care about with your name attached to it, because you would look like a moron. But yet, if he just floods the internet with a bunch of I'm the best, I'm the best, it shows up, and then whoever wrote this effing post wrote it and just slapped his name in, and you're like, now this looks like crap! So yeah, it's… don't get me…

Slobodan Manić: This came out two days ago, this is new, the post. But to be fair to them, the post is literally "as determined by 5 LLMs." Like, that's the list. They don't care about this, this is purely AI-generated, which is not really an excuse, but at least…

Wil Reynolds: Not an excuse.

Slobodan Manić: I agree, I agree. Don't give them publicity, I absolutely agree with that.

Wil Reynolds: I went apeshit for it on LinkedIn, because I got tagged in on that low-quality shit, and now all my notifications are like, I'm on the list, like, like, like. And it's like, see, here's the thing. We're so fucking, like… our egos and shit are so fragile, that if somebody says… so you take a bunch of AI experts, and ask them, and you put them on a list that you go, it was generated by AI, which means they know that those lists are susceptible to all this shit. And they're like, my name on it, like, like, like, like. And you're like, wow! You need likes. That's a bigger problem. You need likes that bad to feel good about yourself. That you're taking something that you know is slop, and you know it's produced with no human input to say, like, I've curated this list, but because it mentions your name, you're like, like. That's the fucking problem, right? It's fucking ridiculous that people would like this and share it on LinkedIn, when it's complete dogshit, and we're supposed to be the people who know that this is shit. We're supposed to be the people leading this freaking industry, and it's like, whoa, this is shit, but if you put my name on it, like, like, like, like, get the fuck out of here.

Slobodan Manić: That's so terrible.

Wil Reynolds: Brutal!

Slobodan Manić: But also, the author is number 3, right?

Wil Reynolds: Of course!

Juliana Jackson: That one was surprising.

Slobodan Manić: Crazy, but this is…

Juliana Jackson: This reminds me of 2015, when we started writing blog posts about tools, and number 3 was the tool, and you were trying to be low-key. That trash in your tool. It was still the top CRO tool, and then it's like, oh, it's us!

Wil Reynolds: You know what's really crazy is if you click through to her profile, it's like an AI image of her, and you're just like, oh, this is just all gangster. Like, this is all suspect. Like, it just looks… it's just a bad look for our industry when people who are supposed to be leading it allow this stuff to proliferate without calling it out for what it is.

Slobodan Manić: Couldn't agree more. This is useless. This is pure trash, like you said.

Wil Reynolds: But, you know, we like trash, I guess, we're trashy.

Slobodan Manić: Not everyone, come on, not everyone. Juliana just typed in chat, she doesn't like being on lists.

Juliana Jackson: Listen, I tell people, stop adding me. Stop adding me, because, like, it's not, like, if it's warranted, listen, like, if it's warranted and curated, then I know the people behind mean good, I don't mind. But if I know this is bullshit so you can get lead generation, fuck off. Like, I'm not… Oh, yeah. Okay, hate me. You can, you know, you can hate me, think I'm a bitch, whatever, you know, like, stay in life. It's cool. Like, I just don't like when we're self-congratulating.

Wil Reynolds: Circle jerk, right?

Juliana Jackson: Yeah, that's what I wanted to say, but I didn't say it.

Slobodan Manić: Hashtags.

Wil Reynolds: Well, you know what? I have a blog… it's funny when you've been writing long enough, I think one of the beauties of writing what you feel and saying what you feel, Juliana, long enough, is you get to look back at the old you to see if you've changed. And, like, I think one of my first four posts was, like, entrepreneurship is a self-congratulatory circle jerk. And that was like… and that has to be 2012-ish. I was like, oh, man, all these people just sit around and being, like, crushing it, and we all know that we're all getting crushed somewhere, and why ain't anybody saying it? And this shit is false. So, I've always been like, this stuff is self-congratulatory bullshit. Like, let's just get the work done and try to do good work. Like, not all this, like, bro bullshit. Anyway, Juliana, I feel like you and I should grab a beer sometime.

Juliana Jackson: I'm bored, like, I'm… yeah.

Wil Reynolds: Let it rip when we're not being recorded.

Slobodan Manić: I love that. Let's… we have 14 more minutes for both of you. What are you seeing out there that others are not talking about yet? Talking about trust in the age of AI? So, we know a lot of things are obvious, getting obvious, becoming obvious. What's the next thing that people will realize was broken?

Kelly Wortham: And Juliana, especially when you think about, like, one of the things that you wrote about was about content, and, like, you had this amazing blog post where you talked about content and images, and how people… and to me, I don't think, especially people in experimentation, I don't think they understand that, and how they can use that, and how they can test that.

Juliana Jackson: Right. Yeah, I mean, listen, I know we all are obsessed with channels and making sure we're attributing and we're chasing people around all the channels possible, but I guess the way I'm thinking about this, the customer journey is super decentralized, it's multi-model, it happens everywhere at once, everywhere, with anybody at once. And I'm saying, if we are stuck still in the world, like, we should… listen, this is not me saying you shouldn't focus on having a good website with good practices and so on, but I'm saying that people don't really come to the website to find out about you. They don't peruse the internet, they come and see your website, and oh my god, I must have this. Unless it's something that really, really triggers you, and I can give one of those examples of this actor that died, and he had this bobblehead that was doing… that was the only thing. But what I'm saying is the only thing that follows people everywhere they are is the creative asset. And the creative asset that we create is kind of the only thing a brand can control. We don't have control as we used to about how we're perceived, about how people use us, about how they understand us, but we still have control over the creative assets. So I think it's important that the creative asset carries the messaging and the positioning and the storytelling that should be relevant across audience, across campaign strategies and goals and so on. So I've been working since last year on a method that I call creative intelligence, and that's kind of how the industry calls it, to understand what is the contribution of the creative asset on the campaign's performance. So, what I'm doing is I'm just breaking the asset apart. Like, I'm decomposing it, like Jim Gianolio would do in MMM, I'm decomposing the asset and all its components, and position, and copy, CTA, subtext, color, whatever, and then I'm measuring how are these things affecting frequency, affecting reach, what is the creative fatigue that will occur at some point? So now I've been focusing on YouTube. I love YouTube, man. Like, for so many reasons, but after all I heard also at Google yesterday, YouTube is gonna be it. It's definitely the tool that people research with, it's definitely the tool that people check out with. So it's very important, I guess, when I'm thinking about the future of the creative brief, or these assets, must have this mix of search intent signals, must have a mix of creative insight signals, behavioral data, all of this stuff needs to go into the creative brief, so they can adapt those assets, because the asset is your storefront, it's your targeting, it's kind of like everything that it is. So, if I was in CRO anymore, I would definitely focus on understanding advertising. Google Ads is gonna really dominate, unfortunately. Like, what I heard… is no escape, so if I was in CRO today, I would learn search, like the new version of search, I would learn ads, and I would try to understand how can I bring that CRO mindset, that experimentation mindset and statistical rigor to analyze the stuff that people actually get in contact with before they make the decision, because that's kind of where you want to be.

Kelly Wortham: Because the decision's happening up funnel, because people are researching long before they get to your site.

Juliana Jackson: And the participation phase, like, that's when you want to be, you want to get people engaged, you want them to participate, and that participation is what permeates afterwards the culture and, you know, the stuff that I put on slide decks at work. So, yeah, that's it.

Kelly Wortham: No, you're absolutely right, and for those of you in the experimentation world that are on the call, there's… people are writing about this outside of the experimentation world, but I'm not seeing people in the experimentation world. I mean, I'm writing about it, go read it, it's on the TLC blog. But I'm writing about it because I'm not seeing it in our industry. So, like, especially you'll… we're gonna start seeing tests that are inconclusive, because as the AI-referred traffic comes in, those people are already persuaded, they're coming in having already been influenced by the research they did, and we are running on our websites tests to persuade, or whatever. Maybe you have a Help Me Choose feature that you want to run a test on. They've already made their choice, and they're just landing on the website, you know, trying to confirm what they already learned.

Juliana Jackson: Exactly.

Kelly Wortham: They don't care about your tests.

Juliana Jackson: It's back to what Wil said. Make sure that the shit that you got on the website fits what people say on Reddit, what people say on, I don't know, YouTube, like, it's kind of the other way around. Before it was, oh, we have to make sure it's as per the website. No, nobody gives a shit. You have to make sure you understand your brand enough, so that when people come, finally, to do that purchase, you can cross-check and cross-reference what you heard.

Kelly Wortham: Standing impatient. But at the same time, it doesn't mean that the beautiful journey that we all created goes away, because the people who didn't come through AI-referred traffic still matter. And the AI needs that beautiful journey. Well, it doesn't have to be beautiful, it has to be… what's the words? Come on, Sani?

Slobodan Manić: Not broken?

Kelly Wortham: Easy for them to scan, machine readable, right?

Slobodan Manić: Flexible. Non-broken, yeah. Non-broken!

Kelly Wortham: For the machine to read. It's all gotta be there, and it's gotta be clear and easy for them to read, so…

Slobodan Manić: I don't want to upset any designers, but typography is not going to sell a product. Colors are not going to sell a product. They never should have sold the products.

Kelly Wortham: Get rid of the carousels. Yeah.

Slobodan Manić: Oh, oh, oh, the bubbles are bursting, okay. But come on, if you're selling at your checkout page, or using the fonts on your website, it worked, I'm not doubting that, but that was the wrong way. You're just, you know, instead of eating better to look healthy, you put on a Facetune filter to look healthier for a split second, like, it shouldn't be like that. Fix the core.

Kelly Wortham: Yep.

Juliana Jackson: I miss spinning wheels, man. That's my next LinkedIn post.

Slobodan Manić: No, no!

Wil Reynolds: Punch a monkey, remember that?

Juliana Jackson: That was so good! Punch the monkey.

Wil Reynolds: Those ads?

Juliana Jackson: Those were better times.

Slobodan Manić: Simpler times, not better.

Wil Reynolds: Simpler.

Kelly Wortham: And just a little note for our next conversation next month, we're gonna talk about the intersectionality of accessibility and SEO and AI discoverability. So, like, how all of these things come together. So, like, how usable is your website? And Elisa will be here, and Shirley Van Hollen is an expert in accessibility, and then, of course, Sani will be back, because…

Slobodan Manić: I want to tease something for that episode. Go to Google Trends and search for web accessibility. There's a baby spike in July 2025, when the European Accessibility Law came to power and people had to get informed. There's a huge spike when coding agents started building websites in February 2026. So, we did not care, nobody cared about… or the people, or anything, but now that AI needs to use websites, more people or non-people care, I don't know who it is that's doing… maybe it's coding agents doing those searches, how to fix that stuff, so… it's insane that we didn't fix this already, and it just bothers me.

Kelly Wortham: We don't care about the people, but we care about the robots!

Wil Reynolds: One thing I will add here to be very careful of is what's in your footer, because it gets repeated. So I've seen a lot of banks, specifically banks, where they've been made by legal to put certain things in their footer related to, like, rates or something like that. But then that rate isn't just on… that rate now exists in very easy to indexable text, very often repeated over and over and over again on a bunch of different pages on your site, which can cause your wrong rates to show up, which I thought was very interesting, because now you have a rate in a table that's hard for it to read, but you've got this nice, clean block of text at the bottom, which then leads to customers thinking you have rates you don't, which then means they start to call your salespeople and spend a lot of their time, and then when the salesperson goes, I don't know where you got that information. They're now angry that that's not true. The other thing that's very interesting is if you're launching a business or anything, how you name it matters more than ever. So, I learned a lot working in banks, but they all are like, First Republic this, and now they're all going, why are all of my rates wrong? And it's like, well, you named your bank the same name as 8 other banks that have the same name in other states, so how is the AI supposed to be so smart that it's supposed to know which one to apply to which one? And now you have the wrong rate showing up, and now you've got upset customers. So, like, even, like, how you name your businesses, your organizations, your products, your services, is going to have major long-term implications. So it's something to think about now if you're in a redesign or a relaunch. Like, somebody's got to speak up about that, because you will be creating tech debt that you cannot unfuck once everybody's already said, this is the name of the brand now, good luck!

Slobodan Manić: I did 3 audits for 3 companies that had the same problem. Let's say they're… I don't know, let's say they're a gardening company, they're not, but let's say they're gardening, and the company name was, We Make Gardens Better, or something stupid like that. Good luck getting picked by the LLM and recognized as an entity if you're so generic. And there's no good solution, the only good solution is really… maybe your owner can have an online persona, so you attach yourself to that. Like, Wil, we talked about the Supabase phenomenon, really, and this is kind of similar, because completely unknown, and I talked about this at a conference two weeks ago, Juliana was there. Like, they built their own identity for agents. They literally did that. They're a Postgres development platform, so a vibe coder doesn't know what that is, but a vibe coder's coding agent knows what that is, and that it needs exactly that. Like, this is… there are games and there are levels of this that are just playing out right now for the first time ever. And it's going to get crazier and crazier and crazier. So, choose your name wisely, choose how you define your entity wisely, and what you attach it to. I think that's even more important. Like, are we in this part of the LLM? Are we in this corner? We never had these problems before. This is completely new.

Wil Reynolds: That's been one of my most fun things. It's only when we started realizing, like, way…

Slobodan Manić: We had a moment that, yeah, exactly.

Wil Reynolds: Yeah, it's like, you know, so, like, start thinking a lot about, if you're in SaaS, think a lot about your documentation. Because, like, you know, you make it… imagine if your pitch is, I make your tokens more efficient, because you don't have to crawl my site and try to figure out what's the right… oh, no, but we got this old code that's sitting out. Even changing your nav? If you are a SaaS company, changing your nav now affects my AI's ability to properly code, because it's like, go here to get your API key, go here to do this. You're like, oh, we made this change in our nav. Well, now, it's telling everybody to go to a place that is wrong, which means I gotta either spend more time to figure out why it's not getting it, screenshot again, being like, which of these navigation items should I go to, or I gotta burn tokens while it's clicking on a bunch of freaking things, reading all these pages, trying to figure it out. So, just something else to think a lot about is the name changing, the navigation changes. If you've been consistent long enough, that shit is burned into the training data, so when you go to change a nav inside of your documentation, or inside of your product, just know that you might end up having to do a lot of work to make sure these agents get it right. So then, talk about it publicly, and be like, we changed our nav! Everybody, it's over here now, right.

Slobodan Manić: I just released an episode with Matt Biilmann from Netlify, the CEO of Netlify, and they have this… he pushed the term agent experience, building a website so agents can experience, not… they went from user experience to developer experience, now agent experience, like, the new term. The entire website should be instructions for agents, like, if the agent can do stuff independently, provision new accounts, all that stuff. This is just becoming a new thing right now, but there are early winners already. Supabase went from $200 million valuation to almost closing a round for $10 billion right now, in a year, because of vibe coding, because they were the default choice.

Kelly Wortham: But is it gonna be bifurcated? Are you gonna have AX and UX?

Slobodan Manić: They can work at the same time, but I think AX is more important than UX because it's the more difficult one. You can do AX and then add UX on top of that. The other way around is more difficult. That's just my take. But you should have both, of course.

Kelly Wortham: Oof, oof, oof. Alright. It is that time. I think it is very interesting that we leave it on that note. AX is more important than UX, considering our very next conversation is going to be about the intersectionality of accessibility, which is about the human, and SEO and AI discoverability. So, hey, let's have that debate next time, shall we? Alright. Thank you all.

Wil Reynolds: Thanks, everybody! Bye. Anybody else?

Kelly Wortham: Have a great weekend!