222: AI Visibility Is a Vanity Metric with Wil Reynolds

Wil Reynolds, founder of Seer Interactive, shares how losing 80% of organic traffic actually revealed that his team had been tracking the wrong metrics for years. We get into why AI visibility is a vanity metric, how 44% of LLM users include brand names in their prompts, the real security risks of wrong phone numbers in AI answers, and why trust (not visibility) is the only moat that matters.

Chapters

• 00:00 - 80% traffic drop, pipeline went up
• 06:43 - The pressure of experimental channels with real budgets
• 08:28 - AI visibility is a vanity metric
• 13:49 - People are tracking the wrong prompts
• 17:15 - AI is getting your phone number wrong (and scammers know it)
• 22:30 - Trust vs visibility: the real optimization target
• 30:26 - Trust is a moat, hacks mortgage your brand
• 32:20 - Be seen, be believed, be chosen
• 44:15 - What digital marketers should do right now
• 49:33 - Where to find Wil

Key Takeaways

1. Traffic is no longer the leading indicator - Seer lost 80% of their organic traffic over two years. Pipeline went up. Social converts at 5x organic. The correlation between search traffic and revenue has broken for many businesses
2. AI visibility is a vanity metric - When ChatGPT doubles the length of an answer, your "visibility" goes up without any more humans seeing you. If visibility grows but leads don't, you're the sucker. Track visibility against your pipeline, not in isolation
3. 44% of LLM users put brand names in their prompts - Wil's team watched real humans use LLMs and found nearly half include specific brands. Head-to-head brand comparisons are the prompts worth tracking, not generic category queries
4. Trust is the only real moat - Michelin built a restaurant guide in 1900 that still drives foot traffic and pricing power. RAMP launched 52 AI-generated restaurant pages in a day. One is rotisserie chicken made by a chef. The other is a chicken nugget. Both are chicken, but only one builds trust
5. Be seen, be believed, be chosen - Visibility gets you in the room. But if people Google your team and nobody shares their content, if you're not speaking at conferences, if your newsletter has no engaged readers, belief falls apart. Trust transfers from people, not listicles

Concepts Discussed

• Three Types of AI Search | Search-led (web index + AI), answer-led (hybrid with tool use), and fully generative (training data only). Each requires different optimization.
• Training Data Lag | Gemini 3.1 launched with training data from January 2025, a 16-month gap. Work done since then has zero provable impact on training-data-based answers.
• Brand-in-Prompt Behavior | 44% of observed LLM users include brand names in their prompts. Changes the optimization target from "show up for generic queries" to "win head-to-head comparisons".
• Phone Number Hallucination | LLMs serve wrong phone numbers for businesses, creating fraud exposure. Especially dangerous for financial services targeting elderly customers.
• Recommended Stack Visibility | AI coding tools recommend tech stacks from training data. Being the default recommendation in Claude Code or Codex creates durable, hard-to-displace visibility.

Connect with Wil Reynolds

• LinkedIn: linkedin.com/in/wilreynolds
• LinkedIn Newsletter: Thinking Out Loud
• Seer Interactive: seerinteractive.com

No Hacks is a podcast about web performance, technical SEO, and the agentic web. Hosted by Slobodan "Sani" Manic.

Transcript

Wil Reynolds
===

[00:00:00] Sani: Wil, you recently wrote about your organic traffic dropping by 39% last year, but the revenue held up. What was the point where you realized it seems broken, but it's not really broken, or it always wa was broken? Like what? What was the thought process?

[00:00:17] Wil: Oh man. So I gotta give people the full story, so it might take a little while for me to get it out. I had a team of four people on marketing. So actually the number FYI over two years is now like 75, 80%. That's organic traffic drop. That's it right there. Like it's legit. So when I see these reports that are like, Google's not, I'm like, I don't care what anybody says.

[00:00:35] Wil: I go into my own analytics. Now tell me how Google's traffic's not really down. Get the freak outta here. I lost 80% of my traffic dog I had three and a half people on my marketing team. And after the pandemic I think Seer is a really safe place to be during uncertainty because people know that I'm not gonna just wake up and start firing people because the economy's bad.

[00:00:54] Wil: But once the pan, so everybody stayed during the pandemic, we didn't let anybody go. And then once it got safe, you start seeing some people being like, ah, it's time for me to move on. It's safer out there, whatever. So I went from four, from three and a half people to three to two, to one to zero on my marketing.

[00:01:11] Wil: Over a course of maybe I don't know, maybe like a year and a half. So then I'm sitting there saying, crystal, when am I gonna get some resources to backfill? I went from three to two to one. Am I gonna get one? Can I get half? She's nah. So what ends up happening is I go to zero freaking people on marketing, and I'm running marketing.

[00:01:28] Wil: I'm the guy sending out the freaking company tweets and all that shit. It instantly put into focus that I was putting my team on the wrong metrics. First thing happened because when I took over, I really started focusing a lot on social and I realized that my social traffic converts at a five x over my organic traffic.

[00:01:47] Wil: So I'm like wait a second. And I get the human signal that what I've done created value for people. So I watched my traffic just tank. Then I overlaid my my, my pipeline and I'm like wait a second. Pipeline's up, traffic's down. Maybe this is no longer the leading indicator that I thought it was.

[00:02:10] Wil: And then I went what else is up? And then I went, social's up, directs up, branded search is up. And I'm like, okay, there's something going on here.

[00:02:19] Sani: Did it break last year, two years ago, or was it broken for a long time and most people just didn't notice this model?

[00:02:27] Wil: No, it, you can literally see I, I have my pipeline tracked against my search traffic and they followed a very similar trajectory and then all of a sudden one day you see the bifurcation, it keeps going down, but then the pipeline went up and I'm like we are at an inflection moment right at that point.

[00:02:45] Wil: And it just so happened to be around the same time that ChatGPT started taking off

[00:02:50] Sani: And then Google ai mode and overviews and all that.

[00:02:53] Wil: Yep. Barred or whatever they called it

[00:02:56] Sani: Way back then. Yeah, exactly. You and your team wrote an article about G-E-O-G-E-O is a hot term right now, and everyone has their thoughts on them. If you wanna share your thoughts on GEO as a, as just an umbrella term for whatever's happening post SEO, you can share that.

[00:03:10] Sani: But you wrote about it and you, there are three different types of AI agencies, what you said. There's search led, answer led, and there's fully generative AI engine as well. Each one of them should be treated differently, is what you claim there. What are the main differences for web professionals who are trying to build their careers and sustain their careers in 2026?

[00:03:30] Wil: So we've gotta update that one a little bit.

[00:03:32] Wil: We owe the world an update, but at the time we were like, we is no, just AI search. It was like, at the time you had like perplexity, which at the time was like one of the early movers and you're like, that's like a web index plus ai. And then you had the hybrids at that time you had no tool usage from, you had no tool usage from Claude, none, no web access.

[00:03:59] Wil: So then you had these other ones that were like purely training. And then you had the ones that were starting to use web as a tool and they could go out. And it meant that those three things were different optimization paths. And you saw people just saying, and you know what the thing I find is.

[00:04:16] Wil: When someone internally is course correcting us, we're like, oh, snap. We got this wrong. And it was one of our guys, this guy Jason, he's I'm watching you guys talking about optimizing ai. He's a developer. He's like, all you guys are not thinking about this the right way. And he recorded an internal video, broke it all down.

[00:04:35] Wil: I transcribed the video, turned it into the turned it into the post and was like, thank you Jason, for course correcting us. I think when something's new. We all need to take that step back and be open to being wrong. I think we're at a moment now where all of us need to be open to being wrong more often than we were two or three years ago, because the only learning comes from putting your stuff out there and then having other people challenge it, and then sometimes being like, you're right.

[00:05:05] Sani: A hundred percent. I had a friend on this podcast. Maybe you know him, Brent Res. He's a. Yes he's a very close friend of mine, and what he said on the podcast was, this is not year 26 of digital marketing and the internet. This is year one of something new. Like whatever that is and what you just said, let's make the mistakes, let's learn let's just figure it out, is the right approach right now.

[00:05:27] Sani: Because if someone tells you they know exactly how AI search works, I think they might be lying to you.

[00:05:35] Wil: I've got quotes from Dario Amede and Sam Altman saying that they don't know how their own AI answer questions. So if the people that are the CEOs of these companies are literally saying, I don't know how this thing answers questions, enough to answer it for you. Then what is some guy like me saying, I know how this stuff works.

[00:05:56] Wil: No, I think Brent's totally right. The difference I think is when we were all screwing around with search engines back in the day, it was an experimental channel, so the pressure was low.

[00:06:14] Sani: Yeah.

[00:06:16] Wil: So everybody that bought search at that time knew what are these kids doing over here? Maybe this search engine thing will take off.

[00:06:23] Wil: Let's throw 'em a little, let's throw 'em a couple thousand dollars and see what happens. Now you've got search, search as a major investment. People are watching it not perform as well, and they're applying a pressure on an experimental channel, and they don't want to deal with the reality that we are all learning now.

[00:06:43] Wil: It was easier when search engines launched. It was like, oh, we're all learning. You're like, you know what? How much of my budget are you? 0.5%. No problem. Now, when you're 20, 30, 40% of someone's budget and it starts to crater, they don't have the patience that they need. I don't give a fuck what you say. I can't call Sundar and be like, yo, can you update the training model?

[00:07:09] Wil: My client's upset, and this is the examples that I use when I talk to people, like the answers that come outta training data. Gemini three, one launched what, like a month ago. Its training data is from January 1st, 2025. So now. All the work that every geo expert all of us have done since January of 2025.

[00:07:36] Wil: All the prompts that are answered by training data, you have no idea if that work is helping or hurting. It is a 16 month pause. It's an experiment you did 16 months ago that you have no idea whether or not it's working or not.

[00:07:52] Sani: And we had a conversation on LinkedIn about building tools and versions and recommended, we'll talk about that a bit later. Another post you had was about AI overviews, and I watched the video that you recorded. It was two months ago. You did a search, it gave you 17 brand links for time tracking software.

[00:08:08] Sani: I believe it was 15 of them opened a Google search. What does that tell us about the state of click and zero click marketing and then, and like all that stuff, if you don't even getting that click within the box that's supposed to hide clicks away from, what does that mean for the future?

[00:08:28] Wil: So here you go. Anyone who is tracking, it's so hard, man. We're at such an early phase If you are tracking traffic, no, let's start with visibility. You gotta first be visible before you can get traffic. If you're tracking visibility. And you don't also track things like the number of words or brands mentioned per model per prompt.

[00:08:52] Wil: Over time, you would not know that back in November, chat, GPT over doubled the length of the answer.

[00:09:01] Sani: Yep.

[00:09:02] Wil: So your visibility went up. Did the register, did the cash register go up? No. It's the same amount of people searching. They just put a lot more of the answer in. So that's a chat chip t example on Google.

[00:09:13] Wil: These people are now take, we've been mentally trained for 20 plus years that if you see a link under a company and you click it, it will take you to the company's website. Now in the AI overviews, if you click it, it takes you to a Google search for their brand. Another one of these studies that come out, Google search is not down, and you're like, did you know this person Never responded to me. You're like, come on, dog. Like stop being Mr. Right guy on everything. We're all learning. So stop trying to act like you didn't see this like you came out with a study that did not look at this, that's okay 'cause we're all learning. Thank you for putting it out. But now you gotta come back when somebody gives you new information and be like, oh, maybe the reason why Google search is up is because Google has taken the blue links and the AI overviews put it all at the top and made it trigger new branded search.

[00:10:11] Wil: So of course their traffic is up because every single link I click on now in the AI overview runs a Google search. That's an easy way to make your traffic go up, isn't it?

[00:10:19] Sani: Also, it's a Google self-reported metric that works for them. The higher it is, the better it works for them. Like when they launched the nano banana in Google work, like if you have a paid Google account you can use nano banana and every time you go to your email, to your documents, there's a prompt that says, use nano banana.

[00:10:35] Sani: I wrote, I had a LinkedIn post then. Give them a few months to report on how many people have tried and use this, and sure they did. They said millions and millions of users are using nano banana. No, they just clicked once and created a stupid, or was it V or whatever it was? That's not a user, that's a curious person who had nothing better to do that that, that should not count.

[00:10:57] Wil: Monthly actives would be nice. Don't tell me how many people used it. How many people

[00:11:00] Sani: Weekly would be even nicer.

[00:11:02] Wil: Weekly would be even nicer. I

[00:11:03] Sani: There was never a better time to sell vanity metrics than today because there, there's so many of them and it's easy and a lot of companies are doing it, and I think it's the right thing. No. It's a very wrong

[00:11:11] Wil: so true. No. I've never heard somebody say that, but you're

[00:11:14] Sani: It was never easier to sell vanity metrics, and I hope this stage doesn't last very long. had a whole article.

[00:11:21] Wil: shows a history shows it won't.

[00:11:23] Sani: You had a whole article about AI visibility being a vantage metric. That's a hot, I don't want this to be a hot take, but if you look at what most people are saying, especially on LinkedIn, that is a hot take. I don't think it is, but ask the majority, just do the poll and it's a hot take.

[00:11:42] Sani: Why are you so certain about this being the wrong metric to track?

[00:11:48] Wil: I've been around search for a while it's like we're, we've been around for a while Sonny? So it's we know that the vanity metric early was rankings and then people went, wait, I gotta get traffic from those rankings, right? And then I need that traffic to turn in a business or otherwise, what do I have, like sophisticated buyers, right?

[00:12:07] Wil: So to me it's just a regurgitation of what we did years ago, which was. Hey I don't know, like you can be visible. That's great. But then somebody's gotta actually take an action for you to make any money from that visibility. And if you don't track those two metrics against each other, you're the sucker.

[00:12:27] Wil: Because if your visibility is growing. But your leads aren't growing and or if your MQL and SQLs and conversions aren't growing, then it's easy for your visibility to grow. When chat GPT doubles the length of an answer, and it's that didn't get, you didn't get double the amount. It's the same person seeing more answers, right?

[00:12:51] Wil: In the same way that people had to learn that ranking number 30 meant nothing. It's once they list 16, 17 different answers to a problem, you're the human brain's only going to go to four or five of them and make a decision. And because to what Rand's point was, because they refresh so much and so often and they look so different every time, you gotta create that offsetting metric to make sure you don't fall for the vanity

[00:13:14] Sani: What was that joke? The best place to bury a body is page two of Google search, like the years. I think now it's the second paragraph of an AI overview or whatever it is. Because if they're making it longer, like you said, and interestingly enough they announced that change right before they announced ads was, it was tied to that.

[00:13:31] Sani: And you wrote about this as well yet, and this is from your LinkedIn post again, you said that 90% of the questions you get are about AI visibility. So it's

[00:13:42] Wil: is true.

[00:13:43] Sani: people care about this. You also wrote, people are tracking the wrong prompts. What do you mean by that?

[00:13:49] Wil: So we actually watch people go to LLMs so we can see what real people do. Average. SEO wants to spend no time looking at real human behaviors. They want to use all these proxies for humans, but God forbid you go, we need you to actually have a watch a human do this well, oh, I'm gonna build this tool and scrape this thing.

[00:14:14] Wil: No, watch a human. They're the ones that actually take out their wallets. They're the ones that take out their wallets and then put in their credit cards. Semrush doesn't do that. Google doesn't do that. Whatever prompt tracker you have doesn't do that. So it was in watching the humans that we started realizing for one of our clients, for 44% of all the people that we were able to watch them try to solve a problem using an LLM.

[00:14:37] Wil: 44% of them put brands in their prompt. And when I saw that, I went, what I watched the first. 50 seconds of a video and saw somebody put in Deloitte. What did Deloitte and what did another one of the big consulting firms think about this topic instantly paused the screen, ran over to one of my UX people and was like, can you count the amount of time somebody put a brand in?

[00:15:01] Sani: Wow.

[00:15:04] Wil: Visibility matters most when you don't have a brand. If you have a brand, there's a chunk of the market that's putting your name in. So what I think we should be tracking are things not just what are seer interactive strengths and weaknesses? We're tracking those, right?

[00:15:22] Wil: But seeing people do these searches made me go, wait. The other thing I think we've gotta track is these, like head-to-head searches. I, 'cause I think most people, once you realize that listicles are out there, you're going to be like, oh, this is crap. So therefore, like I can't just take the answers they give me because they're so easily influenced that I don't want those answers.

[00:15:43] Wil: What I do want is to go to my WhatsApp groups, LinkedIn groups, LinkedIn friends, text, email, those places, the hard to track places. And say to my friends, Hey, this is my thing. Who do you recommend? Then I'm gonna take those names and brands I was given, and then I'm gonna throw them into the LLM and compare those people against each other based on what I need.

[00:16:08] Wil: And like to me, that is something that not enough of us are tracking. And I built a tool recently and put it on GitHub to be like, you guys should start tracking these comparisons of yourself against other folks.

[00:16:18] Sani: That's great advice. And like I said, it's 44% of all those prompts had a brand in them, so people are coming to those LLMs already having an idea of what they want. It's not fully like starting research from scratch. Also what you mentioned, and this was like the, what, last year? I predicted dark social is going to be the next internet because everything else is flooded with bots and slop and everything.

[00:16:40] Sani: We need to go to those close communities and everything you just said that, that's where the magic is happening. More WhatsApp groups than I'm, than WhatsApp one on one conversations these days and that's where stuff is happening. I agree with that. Another LinkedIn post you had, you have a lot of people you have to follow will on LinkedIn.

[00:16:59] Sani: Like his LinkedIn is just. Every post is amazing. You wrote about most businesses having their phone information wrong or fetched. Explain that please, because this is something I've been obsessing over as I'm working with clients, how easy it is for this to be completely wrong.

[00:17:15] Wil: Sure.

[00:17:16] Wil: A lot of our clients have phone numbers for people to call in. We have some clients that are in banking and finance and other places like that. And somebody in our company, some smart person, started tracking the phone number as a prompt to track and. We, no, but the catalyst for this was, I read a blog post somewhere where somebody was showing how different banks' numbers were showing up wrong, which then created an incentive for spammers to register those numbers to then rip people off.

[00:17:48] Wil: So a scammer now could literally be like, I'm gonna run these prompts against all these major financial and banking institutions. It'd be really good for ones that are like that deal with elderly people that are easier to rip off, right? And now you see when the number doesn't match the actual number on the website.

[00:18:06] Wil: And if it's a completely made up number, and sometimes it'll be like a number for the wrong DEP division or wrong department, but if it's a completely made up number, guess what you can do? Get the number, run your scam. And now you're just clearing out people's bank accounts like crazy. So then we started tracking phone numbers for some of our clients who would be exposed to that because we've gotta protect them and say your, all your tracking numbers on your website is now a problem.

[00:18:31] Wil: All the all the times where you don't reinforce the same number over and over again 'cause you wanna know what part of the site they were on. So you can track phone calls could now result in, and this is the problem with visibility and ai. If my only job is to help you with visibility. Then I should not help you with that problem.

[00:18:49] Sani: That's a

[00:18:49] Wil: That's the problem. It's a bad KPI, but I could save you millions of dollars in legal. And having to refund people their fricking money back. I could pay for my contract 10 times over by helping you have the same phone number show up. But you don't incentivize me for that 'cause you want some fucking vanity metrics that show that you're showing up more.

[00:19:09] Wil: And all I gotta do is sit around and wait for chat GPT to double the results and send you a. Port that month and say, look at all the work we did. Even though Gemini, like I said before, hasn't updated in 16 months. So the training data hasn't updated for a set of your prompts. Like this is the reality that we have to talk to people about and it's why I'm passionate about it.

[00:19:25] Wil: 'cause I hate seeing people get ripped off or focus on the wrong stuff. Only to find out six months later, damn, this didn't work.

[00:19:32] Sani: And we'll see more and more of that. You are in SEO. I've been in SEO long enough, and I have, but not as long local SEO when that was like the most important thing in the world. Ap name, address, phone, that had to be consistent everywhere, or Google doesn't like you. It doesn't accept your entity, and doesn't want to deal with you in any single way.

[00:19:52] Sani: That times a million is what's happening now, because it's not just Google, it's a lot of other systems that if your number is not consistent, number one, you can get into legal trouble and. You don't want to get into legal trouble, be over something as stupid as that. But number two, why would a machine I talked to Dwayne Forrester about this, and he has his book The Machine Layer and talks about trust a lot, and we'll talk about trust in a second.

[00:20:14] Sani: Why would an AI system choose your entity that's completely inconsistent at all over the place? Why not just pick your competitor?

[00:20:22] Wil: So I'm gonna butcher this because I'm learning still, but I read an article recently, it was a research study that got into something that was like, the more your company sounds like other companies. The harder it is for LLMs to attach the distinct factors to any of you, because it's if 90% of the pages we're in industry leading this and da, and we work with this, so now you gotta think for these banks.

[00:20:53] Wil: It's if your page all kind of sound the same. It could be a problem that it starts confusing whose number is whose for your, because all this other stuff sounds just like everybody else. Sounds so now I'm just gonna predict that your phone number could be one of these other numbers. So yeah, it's man I'm, man, I, it's a crazy freaking time.

[00:21:10] Wil: It's a crazy time to think about stuff like that.

[00:21:13] Sani: Yeah, because they just associate things with different, with other things and then if your bank. That's a really good example. I, a friend of mine told me about this. It's a known thing with lms. Start a sentence and you say, this person likes color yellow. What is their job? And it will say, beekeeper school bus driver. And I forgot what the third one was, because it associates color yellow with anything that's yellow. Yeah. That's how stupid. Not stupid or smart. That's how LMS work because they associate things.

[00:21:45] Wil: it's stupid. It's stupid. It's I think we give them way too much credit. I think the problem is we say stupid because we give them so much credit. 'cause they're so good at certain things. And it's almost it's almost like taking somebody who's an amazing sumo wrestler and putting them on like the pommel horse and being like why?

[00:22:01] Wil: Why aren't they good at all the things? And it's like you. You. You are an athlete, so you really have to be able to, I think it's so helpful. I think that example, Soni alone is such a good example for people because it helps them realize, oh, all of this credit that I'm giving this thing that is amazing at these things, I'm now getting lazy and allowing myself to attribute that same strength in areas where it's not strong at.

[00:22:29] Wil: So I think it's a great

[00:22:30] Sani: tested that with Claude Gemini and Chad GPTA few months ago. Exactly the same. So this is something to think about because they just associate entities and predict what to say next based on what, what came before. It's wild. So let's talk about trust versus visibility. Trust is it is the most important thing a course, if you ask me.

[00:22:53] Sani: It's the most important thing online, whether you're a personal brand, whether your company, how do you optimize for trust online?

[00:23:02] Wil: All right. I'll tell you a story.

[00:23:05] Wil: We did an experiment where we had our footers said something like Philadelphia founded remote, first company, whatever. We got rid of that and then put in our company retention rate where like we have 130 clients, 97% retention rate is what it was last year, three months ago, we found out that it went down from 97 to 92, we updated the footer.

[00:23:25] Wil: That's how you earn trust. You've gotta say Hey, I'm not, I posted these numbers about my company when they were good, then they went down and I'm still gonna post them. I owe it to you to try to get these LLMs real information. And then what I think happens is agents are gonna be interesting in the sense of I will still spend some amount of time looking at a travel charger.

[00:23:49] Wil: I will not just let an agent just run off and buy me one. So you say, of course. But when we think about agentic browsing and how hot people are on like agents are gonna do these things, I'm like, I still think that humans like to see and touch and feel certain things more than other things.

[00:24:08] Wil: Some of us will let a rando person buy our groceries. I'm not one of those people 'cause I actually want to feel my stuff. I like to make decisions about those things. Some people have no problem with that, so there are some people that'll travel charger and let it rip. When it comes to picking an agency or picking a service, I have this weird feeling that humans still want to be in the loop.

[00:24:33] Wil: I think what's gonna happen is when you get an answer from an LLM, you're still using all the stuff up here. What did my friend say? Have I seen that brand? Have I seen that brand before? If you type in geo agencies on Google and you get the answer box, I think I've only heard of two of the 12 companies that showed up.

[00:24:56] Wil: I've been in this industry for 20. I started in 1999, so it's 27 years and I have not heard of 90% of the people that you put in front of me. One of them is reputable and the other one that I heard of was like, nah, they just do like listicles. I've heard of them 'cause they're in my region.

[00:25:16] Wil: That's crazy. And I think that this idea of yes, I'm going to trust, but then verify is the key like. That's why I get back to that concept of you can be visible and be seen, but then people want to figure out whether or not they believe you. So now you're trying to get into the content that you wrote and they go, does that make me, even what you just said, Sonny?

[00:25:36] Wil: You're like, Ooh, everybody should follow Will on LinkedIn. Why do you say that? It's a trust transfer. We're all looking for the curators, the people who are gonna say in a world of infinite content, we are now in a world now of infinite content. Our brains have not gotten any different than they were. 300 years ago or whatever, right?

[00:25:57] Wil: So our brains cannot process infinite content, so our brains need shortcuts. The shortcut is people I trust saying you can trust these other people, you should read their stuff, et cetera. And I think so many people are getting their AI visibility with things that diminish their trust. Once you see them that then you're like, so if this is how you got visible in AI, and I can uncover that, then.

[00:26:23] Wil: I wonder how you run the rest of your freaking business.

[00:26:26] Sani: Excellent point. And also these. New age hacks, AI era hacks. And there's a reason the podcast is called No Hacks because I fucking hate that

[00:26:36] Wil: dude. Me

[00:26:37] Sani: they will hurt you way more long term than any the keyword stuffing child's play compared to what? AI reputation management like all those people, and I keep, I say this in every single episode, all those people who suggested go on Reddit and write a thousand posts praising your brand.

[00:26:55] Sani: I don't see them anymore. And the damage they've done to their clients, it's not going to be easy to recover from that because that stays in the lms. Once that gets into the training data that says, this brand is fine, but they have some shady stuff that they try to do and someone discovers them for the first time, they see both of those things. Good luck recovering from

[00:27:17] Wil: Good luck.

[00:27:18] Sani: I'm saying. So

[00:27:19] Wil: I'll tell you, I have not shared this publicly yet, so I'm gonna do it in a presentation on to on

[00:27:25] Wil: But. I'm showing two different companies going after like restaurant like content and one of them is Michelin, and you're like, Michelin started off as content a book in 1900 and to this day, once you get a Michelin star, I broke down the economics of what one star, two stars and three stars does to your foot traffic and your pricing.

[00:27:50] Wil: So like that star. Built trust. They have not just said, Hey, you're in the thing. It also results in real business, right? Then I showed how RAMP built a bunch of content in a day that got the top restaurants in all major states, in all states. So they launched 52 pages in one day. Do we wanna drive around and actually taste this food?

[00:28:17] Wil: Why would we do that? We can just go to an LLM and you're like, God damn. So now when somebody looks at this trust concept, I'm like, both organizations have content about restaurant. If I said when you come to Philly, the place you should go to is this place that I was referred by ramp. And then I showed you, you'd be like, who the freak are they?

[00:28:38] Wil: And I showed you the content you'd be like. I now have less trust in will. It's like we all know the difference between a great rotisserie chicken done by a chef and a chicken nugget, and right now people are making chicken nugget content and being like, but it's chicken. You're like,

[00:28:54] Sani: And that's the episode teaser right there, because that, that soundbite is amazing. Rotisserie chicken

[00:28:59] Wil: but it's

[00:29:00] Sani: It's absolutely true and

[00:29:01] Wil: We're acting like it's all chicken. Oh, content is content. So I'm just gonna. Write content and create a review and create whatever. And I did it instantly. Ooh, thanks boss. I did it instantly. I did it fast and I got a workflow that makes it look like it wasn't ai. I took out all my M dashes.

[00:29:16] Wil: Ooh, thanks. And it's or you could go taste the fucking food and actually put your fucking reputation out there to be like, we are going to go taste the food and tell you what we think. Ugh.

[00:29:29] Sani: By the way, you know the story of the Michelin Guide how, why they created it.

[00:29:32] Wil: Oh yeah, it's genius. Genius, genius.

[00:29:36] Sani: That's the marketing we need that kind of stuff. And not, like you said, the ramp, I'm not aware of ramp. It probably is an American website.

[00:29:45] Wil: Venture backed American

[00:29:47] Sani: If anyone can do that and recreate that in a week, the value of that doesn't exist. Like the whole thing with I launched an app. I created an app. This is my SaaS. I wrote a blog. A million people will do that tomorrow, which means yours doesn't matter. There has to be something uniquely yours.

[00:30:03] Sani: There has to be something. Dwayne has a post about Dwayne Forrester. Content was a moat. Now context is a moat. This is what he wrote about on Sunday like this weekend. And I agree it has to be something that only you can say. It has to be something that only you can contribute and not something that anybody can just create.

[00:30:20] Sani: When I would one or two prompts. 'cause those days are gone. Those days are gone. Yeah. That brings

[00:30:26] Wil: Trust is a

[00:30:27] Sani: trust is a mode. Absolutely. Trust is the greatest

[00:30:29] Wil: a moat and I think you gotta be careful.

[00:30:31] Wil: I think now, see, this is what's, this is to me. I feel like all this fast AI stuff. It's very I'm gonna mortgage this brand's trust. For this quick win and I got my share. So as long as I then get out at some point I'm gonna get mine win, whatever.

[00:30:47] Wil: And I'm not here for the downfall of the crap that happens to the brand when this all goes down and it's oh, that's gonna be somebody else's problem. And it was interesting. RAMP is a B2B finance company, right? So it's like they're not even a restaurant, but neither Michelin makes tires. So it's oh, like they made tires and made it work.

[00:31:05] Wil: We can make it work. The saddest part. The saddest part about all this hacky stuff is I found this example, went to their website. I'm now getting bombarded by their sales team

[00:31:16] Sani: Of course.

[00:31:17] Wil: first thing. So now I'm getting the, Hey, guess. Guess you didn't see my last five emails. I saw all five. I'm actually shredding you publicly in a presentation, but you don't know that because your hacky tools are being like C level at, I looked you up in clay and now it's telling me that you're in market.

[00:31:32] Wil: It's okay. And then I emailed my CFO to be like, Hey, I saw some stuff from this company for the product side that was interesting. And he's I hate them. He's I don't like them. Their salespeople are super pushy. So then I deleted the email and now I'm like, okay, I guess it's not worth the headache.

[00:31:50] Wil: Their past behaviors affected my CFO's trust. So when I saw their product stack and saw some interesting things for the business, he was like, please don't make me have to work with these people. They pound the phones, they call me on my cell phone, they find my cell phone number somehow. He's please, no.

[00:32:07] Wil: So here they are getting visibility, but if people's visibility is we don't trust you, we don't believe in you, then what are you actually doing

[00:32:13] Sani: The wrong things. Be seen, be believed, be chosen. This is Alisa your colleague. That's her thing. You told me about

[00:32:20] Wil: that's her

[00:32:20] Sani: But you posted about this as well. It's a, it's an AI optimization playbook. That's what you called it in LinkedIn post. Let's go through it. Step by step, be seen. That's the right kind of visibility, I assume the thing we talked about, but not in a hacky way, pushy way.

[00:32:34] Sani: Can you,

[00:32:35] Wil: It is all visibility. I, 'cause you know what it's like, with, I think whenever you're in search, there's always this sense of Ooh, holier than th So I'm like, I don't care how you can spam the hell outta the LLMs if you want, as long as you disclose it to your client and your client knows what risks they're taking.

[00:32:49] Wil: And the immortal words of our old Pope, our former Pope, who am I to judge? But being seen is about getting visible. So I understand that in order to make money, you have to be seen somewhere, whether it's in chats, whatever, but you need to be seen somewhere. However you get that visibility is up to you.

[00:33:10] Wil: So I understand the need for tracking visibility, but then that visibility has to lead to an action that gets the person closer. To the cash register and I think what happens with brands is okay, I started to see you, but then when I read your content, you might've used content to be visible.

[00:33:30] Wil: That was, eh. When I go to read your blog or sign up for your newsletter, and I realize that you have nothing that makes me think, I now don't believe that you are a top geo agency, which is what your visibility told me. Your visibility said you were at the top. But now when I actually assess, hey, are you gonna be at this conference this year?

[00:33:50] Wil: No, this one, that one I'm going to no. If you were at the top, they would want you to speak at those things, right? Now let me run your, your, let me look at your LinkedIn profile since you're a top thought leader in this space. Ooh. Nobody shares any of your leaders' content. So your top three people in geo post stuff and get two likes, and one of 'em is from your CEO and the other one's from that person's mom.

[00:34:11] Wil: That doesn't really sound like thought leadership, just because you put thought leadership in 15 times and did a listicle, doesn't actually become that. So now I don't believe you. I think the bo, I think the breaking in belief is where people go out to their network and go, I saw these companies, but I don't believe these companies.

[00:34:29] Wil: And that's when they go out to social, WhatsApp, whatever, ask their friends. And then I think they go back to LLMs, compare the names they heard. But now it's like that belief part is starting to click. So for us, belief is about do you sign up for my newsletter? It's a real human. I always say this man, and I can't wait to start challenging these punk ass companies.

[00:34:51] Wil: I said, okay, here's how you find whether or not they're doing listicles and in their sales pitch, tell 'em to open Google Analytics in the middle of the fucking call. Show me those listicles and then show me the source medium, because I wanna see whether or not anybody shared it. Does it get direct traffic?

[00:35:11] Wil: Like where does your traffic come from? Oh, it's 90% LLMs. Shocker. Because you found a hack that worked, but actually nobody believes that shit when they get to it because nobody shares it with anybody else.

[00:35:24] Sani: Look all hacks work until they don't. That, that, that's a definition of a hack. It's something that, is not supposed to work today, but somehow it does. And you know it's not going to work in the future. But you don't tell that to your client. Let's talk about your cloud code adventure. Let's talk about our LinkedIn conversation about the cut of dates and let's cover that first because you discovered superb base because you were building an app using cloud code. That's another layer of visibility for companies that is not being discussed.

[00:35:55] Sani: Are you in the recommended stack? This will make, maybe not tech unicorns, but big tech companies if you, if they get into that recommended stack. It does a lot of money to be made. However, it never recommends the current version or the current software or whatever else. That's a bit of a problem.

[00:36:13] Sani: So tell me about the stuff you've been doing and the stuff you've been building with CLO code.

[00:36:18] Wil: Another shout out to Elisa. Something that she's helping me to realize is, oh man, this is a hard one 'cause it's new, so I haven't worked out how to talk about it, but it's like solutions are gonna become more disposable and it's huh? She like dropped that on me the other day and I'm like, shit, you're right.

[00:36:34] Wil: Like I'm listening to a client problem. Building my own thing specific to their problem, and then it doesn't need to be used for every single person after me, which is a very different way of how agencies have worked is you start getting people into Claude Code, they start being better at listening to the client, figuring out exactly what they want.

[00:36:51] Wil: Figuring out how to build it in clawed faster than trying to talk, send in a request to the company, can I get a this, can I get a that, ah, whatever. So the thing that I've been doing is building more and more bespoke tools to solve bespoke problems for my clients. I don't think it's, I don't think you can say to your client anymore you can now because everything's still new.

[00:37:11] Wil: I think it may be two years from now telling your client, oh, that can't be done because of a software limitation. I wouldn't accept that today, but I think it's gonna be a couple years before that's the norm. And your clients are gonna be like, no, can't you just build me what I asked for and then that's gonna be a problem?

[00:37:27] Wil: So the thing is build me, what I asked for when you're a new like me is I just take whatever you recommended as the stack because the job to be done is to solve my client's problem. Not to be like, should I use OAuth or Google's all I don't give a shit. I don't care. Like I need to solve this problem for my client.

[00:37:47] Wil: So what happens is in Claude does not use the web natively. So Claude code is the big thing that a lot of us are using. It does not use the web natively, which means you have to specifically tell it to go to the web and approve it to go to the web, which means it's training data is what it's telling in your stack.

[00:38:04] Wil: So for me, dude. If I wanted to win in LLMs and I had a visionary enough CEO, I'd be like, these things don't update. But once every 16 months or 18 months, whatever they are, you're gonna have no proof that this is working. What I'm telling you is the faster we can get in as a recommended source, it will create a moat for us.

[00:38:25] Wil: For years. For years. 'cause no one else is gonna allow their agency to invest in doing something they can't show the value of in three days, three days, three weeks. Three months at most fired.

[00:38:34] Sani: Monday.

[00:38:35] Wil: So so now I'm like, man, if I ran a company that was part of infrastructure, a super base, a cel, whatever, I would be tracking like crazy, very custom prompts to be like, when somebody's trying to build a tool like this, what's the stack?

[00:38:47] Wil: What's the stack? What's the stack? And then you've got Codex five, two and five three, both of which have been trained slightly differently. So now when they launch new versions of Codex, it's recommending different stacks. And I'm like. Damn, you're tracking dumb prompts. Like I'm looking for a database, I'm looking to create a login.

[00:39:09] Wil: I don't care about none of that. I'm trying to solve a problem. And whatever stack you give me, I go to those websites, sign up for an account, and next thing you know, once I blow through my free credits, once I blow through my number of projects, I'm allowed, I'm now forking out the freaking Amex paying to keep my stuff up and running.

[00:39:24] Wil: So that's what I'm seeing when I'm building things. And I just feel like, so when I go to these geo companies, they're doing tracking. I'm like, you tracking Codex because I got a bunch of SaaS clients that are infrastructure. You tracking Codex? No. Why would we do that? Ah, shit. And then I go to Claude Code up my own one using Open Router and I'm like, I don't need you for this.

[00:39:43] Wil: So yeah, that's what I've been playing

[00:39:44] Sani: Thumbs up for open router. It's absolutely amazing. I love it. I try to use it for everything unless I really cannot, unless it requires, yeah.

[00:39:53] Wil: one thing on Open router. This is crazy. So I'm testing what we're talking about. To see if open router shows up. I went to freaking Codex the other day and was like, Hey, I wanna build something to track this in chat GPT. It was like, oh, here's the chat. G-P-T-A-P-I. And I also think that chat GPT might be a little bit early, only one test early, putting the thumb a bit on search.

[00:40:17] Wil: Our official docs search this, search that by default, that's Codex. But then I was like, oh, I also wanna track Claude and Grok and this. And it started being like, here's all the code you should write for each individual one. I'm like, literally what I was talking about earlier, I'm like, if I was open router, I'd be like, I need to find a company that's thinking this way.

[00:40:39] Wil: And then we are going to do everything to put everything out there that we can to become the default state. 'cause what sucks is it sucks for everybody. If you don't recommend me, open router. Now I have the overhead of managing updates to Claude Deprecation on certain models on all these different ones.

[00:40:55] Wil: 'cause it was like, here's the code for gr, here's the code for Claude, here's the code for chat, GPT 4 0 5 1 5 2. And now I built all this bloated code where I could have one API call. But if I don't know the language as a vibe code or in a new person to this to say open router or give me one that has a single API, then I don't know that.

[00:41:15] Sani: the first place. Like that. That's the number open route could just pay all the YouTubers that do any tenant cloud code content and just say you, you have all the credits you want. Just use us. And in two years, they would be the only company selling tokens via API. That's how it could be It.

[00:41:34] Wil: God, I wish I knew somebody over there 'cause I'd be like if you believed. The other problem is we have venture capital. On the other side of these companies that are looking for quick. So you end up writing low quality shit instead of the stuff that will make this asset a banger over five or six or 10 years.

[00:41:50] Wil: Who knows? It'll still be around. Like it's no quick win. Quick win. Sell it, get it to somebody else. Let it be their problem. But man, man, like that's an, it's a prime example, bro. I was shocked. I went into Codex and I'm like, why are you And because I started being like, I went to Codex and did it.

[00:42:03] Wil: I was like. These guys are searching their own docs first, which is now Ooh, if you start searching your own stuff first, you're recommending your stuff first, which then caused the LLM to believe that it should go to each website and get its content. Instead of saying, no, that whole thing is wrong.

[00:42:18] Wil: So now you're like open ai. You put your thumb on the scale a little bit on the recommendations of using your

[00:42:23] Sani: I wouldn't be shocked if they were,

[00:42:25] Wil: I wouldn't be shocked either.

[00:42:26] Sani: I think super base, CEO or somewhere super for super base said. A few months ago that 90% of their visits on their website are from AI agents, which means that's where they're getting rec consuming document documentation website. Not the main website, documentation website,

[00:42:41] Sani: Which makes all the sense because all of those cloud code requests and all of those Codex requests they go and they look at the documentation and they tell you what to do.

[00:42:48] Sani: So it's a different world . A completely new reality that. Maybe people will talk about in a year or two, but not enough people. We had our chat on LinkedIn and that's the only place where I saw that kind of conversation, like how much this matters. And if you're building a tech unicorn, maybe that's where you should go.

[00:43:07] Sani: Maybe that's what you should focus on.

[00:43:09] Wil: Brother, what do you think about CLIs? So I've been talking to one of our large infrastructure clients are you using your interface to trigger audiences for up and cross sales and things like that. When people log into your interface, like if somebody logs in their interface and starts thinking about going over to the the increase my usage or do this or do that, you could tag people and create audiences around that and then remarket to them.

[00:43:33] Wil: In A CLI world Sani, it's like I don't go to Elle's interface anymore. So now you don't have those abilities to do those things. I've got one of my large infrastructure clients, I drop that video on and they're like, it's gonna take us days to get back to you on this because like we've not thought about the fact that the CLI degrades my marketing signals that used to happen in app.

[00:43:56] Sani: I totally get it. I'm a former web developer, CLI is my preferred way to do anything using a computer. I just, I don't, web UI is just unnecessary. For most things you're trying to do

[00:44:08] Wil: Because you know what you're doing.

[00:44:09] Sani: it's just a layer that's there. I, we can talk about this for days, but I wanna ask you one, I have one more question.

[00:44:15] Sani: One more question. It's March 24th, 2026 as we are recording this. For someone in digital marketing, let's say digital marketing, what should they be doing right now? What should they be changing? What should they stop doing? How should their strategy change with all the things we discussed? The one thing they should drop.

[00:44:39] Wil: Get a Loom account and get really good at recording everything that you do for an entire day, and have the chat tell you what it can build for you that you did that day. That's how you'll know what to drop.

[00:44:54] Wil: Be like, Hey, you did all that. I saw you do all this. Then you opened up Excel, and then you did this, and then you did that, and then you did this, and then you did that. It's here's a script that you can drop into. Whatever tool, set up these three API keys stack, again, being recommended, and I'll build it for you.

[00:45:10] Wil: I'll never have to do it again. Like you got to put the context in. And I think we know how to put context in. We don't. Our brains are making shortcuts the entire time. It's why I like people to leave me voicemails about their problems instead of typing it into a box because you're gonna take shortcuts.

[00:45:29] Wil: I want all that. 

[00:45:30] Sani: So you want

[00:45:30] Wil: I want all that context. So yeah, I think just record what you do, start your day and just hit record and then at the end of your day, upload the file or take the transcript of you talking it out or whatever, and then be like, all right, based on this, between this time and this time I was working on this, build me something that makes it a little bit faster or a little bit better, and tell me why you did it.

[00:45:50] Wil: Tell me the stack and blah, blah, blah, blah, blah. That kind of

[00:45:52] Sani: And look if you had a person shadow you all day and tell you're doing that wrong, you'd get some great results from that. That as well. But you can't pay a person to follow you all day.

[00:46:02] Wil: That's a, it's expensive. It's weird how, the other thing I tell people is, we're in a human business, so I'm like, dude, it shocks me at how few people use AI for coaching.

[00:46:14] Wil: It shocks me at how many people will ask me to Hey, will, can I meet with you so you can help me with this thing?

[00:46:20] Wil: And then it's did you take the transcript from the call and ask Chad PT first? No. It's so why don't you do that first? You might even have better questions for me later. And it also works on your time schedule. Not when I'm available. I'm traveling, I'm flying around, I got soccer practice or whatever.

[00:46:36] Wil: And it really does surprise me. Like I think if anything, being in a company where me and my leadership team are pressing AI as aggressively as we are, it's still to this day, shocks me at how long it takes some people to get, it shocks me and that makes me not very bullish on how fast this is gonna go.

[00:46:55] Wil: Because it brings me back to the Sweiss reflex. GNA Sweiss was the doctor that taught the world that if doctors washed their hands, that they would eliminate germs. And the way he found that out is women were dying after childbirth. And so you know the story.

[00:47:13] Wil: Women are dying after childbirth and he's like watching the different doctors. The one, the women that went to midwives. No crazy spike in deaths. The one that went to doctors in this university hospital were dying, and he realized that the students were working on cadavers in the basement, not cleaning their hands, and went upstairs and then helped deliver babies, and it killed the mothers.

[00:47:32] Wil: You would think that, hey, use some soap would've been accepted by all these doctors because their alternative was killing women. You have a higher likelihood of killing women 'cause you don't want to use this like hard soap and whatnot. But your alternative is, I believe you're killing women doctors, some of the smartest people.

[00:47:54] Wil: They fought him for almost 15 years. 15 years doctors fought him. He went into an insane asylum and then died a month later because he went crazy trying to show the world that you could save lives by just washing your freaking hands. So put that in a 2026 context. You're looking at people like you gotta still feed your family dog.

[00:48:21] Wil: The world is going in this direction. How is it that I see no traces of you doing these things in a company that's this AI forward and is pushing this hard. And some people would say, just fire those people and push 'em out. I'm not that guy, right? I'm like people, it's interesting how pre ai, some of these same people were your champs and they were your winners.

[00:48:42] Wil: And now post ai, I just forgot about you like. That's not the right way to treat people. So I'd rather, I gotta look at myself in the mirror and say I was okay with you before and all the pros you brought, I gotta have some amount of patience with you to get your own aha moment to then get you on this journey.

[00:48:57] Wil: But yeah, man, like I, I'm surprised at how many people are pulling a semi Weis right now, or where I'm like, Dr. Sam Weiss and my leadership team is like Dr. Sowe being like, use this stuff. And then there's a couple people still like slowly, getting into it, it shocks me.

[00:49:11] Sani: I couldn't agree more. I'm not saying automate every part of it. Be humans. Be essentially, stay human please. But man there, there's so many stupid things humans have to do every day and keep doing and repeating that they don't want to do. This was awesome. I'd really enjoyed the entire conversation. What is the best way for the audience to connect with you?

[00:49:33] Wil: LinkedIn. The other thing I'll tell you is I've learned a little bit about LinkedIn's algo. If you. It right now. I have more thoughts. I think you're probably doing this, Sani. I know. I'm doing it. I have more thoughts than LinkedIn is willing to show the people who want to see my thoughts. So what happens is if you post three times in a day, good luck getting that third post visible.

[00:49:53] Wil: So I would tell people to sign up for my newsletter. It's the only way you can pretty much guarantee outside of the algorithm that if Will has three ideas in a day, I'm gonna get to see all three. Because even for my best friends, people like Rand and shit, it will not show me what he's talking about unless I'm like, show me every time he posts.

[00:50:10] Wil: So I'm thinking these days, like follow people on LinkedIn, but also if they have a newsletter, at least for me, it's my bypass around their algorithm not wanting to show. Will had three ideas today, so let's take the one that's most catchy and show it. And it's you're like why am I missing his stuff?

[00:50:25] Wil: And it's because Al goes 

[00:50:28] Sani: what's the URL for the newsletter on LinkedIn. 

[00:50:30] Wil: I don't, it's on LinkedIn. It's, don't even know, bro. I just set it up recently. It's on LinkedIn.

[00:50:34] Sani: Thank you so much. Like I said this was such a great conversation.

[00:50:37] Sani: And to everyone listening, thank you. Thank you for sticking around with this podcast for five plus years, and I will talk to you next week. Take care.

[00:50:45] Oh

[00:51:13] no. Hatch,