Housecall 2026 - Wil Reynolds === Wil Reynolds: [00:00:00] Howdy their friends. Doesn't it feel like we're all playing a big game of whack-a-mole and it's like we're chasing all this stuff around and I wanna be the person to ground you in the fact that our job in marketing is to be seen, to be believed, and to be chosen. Yes, we all are trying to get more visibility. That's great thing, but that visibility has to serve a purpose. It's gotta help us to be believed and then chosen ultimately. That's how monies me. And I think we're all this red pill, blue pill moment for my Matrix fans where we've gotta decide what kind of marketer and marketing do we want to do and are we willing to do to win. And I wanna start off by inspiring you to think about the Michelin Company, A freaking tire company. Y'all. A tire company has become the company that is known about great restaurants. And how did that start? These guys decided that they were gonna build a guide. They were gonna drive all over France, find the top restaurants, and give this guide out to [00:01:00] Chauffeurs. And Chauffeurs, when they were driving families around, would go to these restaurants and they hoped and believed that if that's a big, if their brand was associated with these great memories and these great restaurants, that these folks would remember them and buy their tires from Michelin. Now if we put that in the scene, believed and chosen framework, they built it, tried something experimented and got seen. But today we all know that Michelin star means somewhere. Somebody believes that this restaurant is amazing and when it happens, it's really hard to get a freaking seat there. So they definitely are getting chosen. Now, another company came along and they're in the finance space and they were inspired by this and they said, wait a second. We wanna do something that's outside of our norm, just like how they did that. But it manifested itself way differently. You see, it looks like RAMP started building out content about top restaurants in Minnesota. And when I [00:02:00] saw Minnesota, I was like, damn, dog, Minnesota. You know, Minnesota's a big ass state. It could take seven hours to drive from one side of Minnesota to the other. And I said, oh my God. It just fell into the zombie content trap. They don't want to be like Michelin and actually do the work. They wanna build content for every state and launch it all on the same day because, ugh. And then they're gonna use AI and they're gonna leave all the clues behind, like evolving digital landscape and a bunch of deep research numbers. And then they just said, let's just do it over and over again. That stuff is formulaic, it's slop, and the saddest part of all of it is next quarter. It's gonna be completely ignored because nobody loves that shit. Nobody's gonna send that to their friends. Nobody's like, wow, glad I read this. And there was a decision to be made [00:03:00] here. And the decision that they made was to not drive around and taste food and actually give recommendations. It was use our. Data plus some AI and load me a bunch of pages that we can launch all on the same day and let's just throw some slop out there to the world. And I am here to tell you that I have a better fucking idea. There's a lot of ways that we could use AI that aren't that. So let me show you an example. I've been showing this example for four or five years. If you type in, I want you to stop, put on your marketer hat and think if somebody types in the word black, O-B-G-Y-N, what do you think should be on that landing page? Might there be some black people? Yes, black doctors would be helpful. Um, could it be things related to. The historical context of what happens when black women go to black OBGYNs, that would be helpful. And we got Zocdoc here [00:04:00] with a bunch of white people. Darn it. Zocdoc, right? This is, we have this happening today in our campaigns, and we're paying to send people to pages that don't answer their questions. Let's stop building AI slop and start using AI to solve this problem. And even when I search for other versions of it, I get all these organizations are literally bidding on this, on these words and getting clicks and sending people to a page that has no black women doctors. So we built an AI to solve for this, and I'm gonna tell you the entirety of the stack so that you can take this transcript or whatever and try to build it yourself if you want. We basically said, all right, let's take these search terms, these baskets of words where we're underperforming and paid, and say, all right, what's the landing page that we have? And then we say, here's the persona. You can either make a synthetic persona, you can actually go and do the work. It [00:05:00] doesn't matter. Here's a persona that's likely typing in these words. Go to the page, break my page into sections, and figure out how my title is performing. How's my intro performing? How's my call to action performing given the persona and it will score. The content based on the persona. And then the AI keeps trying to write new body copy, new titles, new descriptions. And you can see in the before and after copy we're trying different things, but this is happening at scale automatically. And then you, then you have a page and you're like, okay, that's 80% of the work. And then we human in the loop. The last part, of course, we're gonna make tweaks and changes at the human level. The other thing I know a lot of you deal with is how do I get more content? And the problem I believe is for so many folks, the people that have the best content are your doctors, your researchers, et cetera. But you can't use up hours of their time to help you write content. So what have we done up till now? We've gotten content marketers who know a little bit about the industry and they kind [00:06:00] like try to take like what's ranking on Google? And no, we thought there was a better way. So we said, why don't we take content on our website, have AI read it. And then if it's old content, have another agent that basically says, okay, I'm gonna go and kick off deep research and see what's different about what's happened or what's changed with this procedure or whatever, between a time and b time. And then what it does is we have, and I'm telling you all the parts of the stack so you know it, so we're using, um, perplexity, sonar, um, to do the deep research and, uh. And then we have, uh, Twilio with is our phone number provider, and VAPI is our, um, VAPI is our ai, uh, voice interface. And what it does is it makes questions based on the gap between the content we have. And what's being researched out there and it calls a doctor or a researcher or whomever, and it gives 'em a phone call. Usually those calls [00:07:00] last between 10 and 15 minutes, and those folks give us all this rich information and then humans go back and kind of tidy it up. That is so much better than what we do today, which is look at who outranks me on Google. Look at what they've done. Try to add in some words and massage some things in, and just do it one step better when you don't stand out at. All that way. This has been amazing because what we have found is, is that time and time again, if you were to take a brand off of a website, the content all sounds the same. And there is research that shows that that leads to LLMs hallucinating more. It is true. It is true. So when we look at the tool that we built our clients, these are 26 plus year people in this company, and they're like, it felt natural. The AI was asking me questions that were sometimes better than what I was being asked by journalists and people that run trade publications. It felt like I was using the internet again for the [00:08:00] first time, and this isn't some spring chicken. This is, Robbie's been at this company for 26 years. He came in skeptical, as does everyone, and they leave impressed. Now, I'm not just gonna say we interviewed people. It helps us with search, it helps us with ai. It's helping us with all these things, right? And the reason why is because we're giving Google and these other, these other LLMs information that other people aren't giving them. And it's coming from a place of expertise, not from a place of, I've gotta build 50 pages at scale. It's coming from experts. And that is the unlock. You can do it with your doctors as well. So, like I was saying before, this is this red pill, blue pill moment where you're gonna be, you're gonna choose to, what kind of marketer do you wanna be? Do you wanna be whack-a-mole marketer who every quarter? I'm just trying to, and I know our, our companies put a lot of pressure on my marketing folks, but we do still have to decide, are we gonna stand up? Are we gonna continue and, and say like, I want to be investing in [00:09:00] something that might take three or four or five years to show value, but I can't have a hundred percent of my portfolio being winning quarter to quarter to quarter to quarter. Because what happens is you end up with spikes and ups and downs and you're just chasing, you're chasing tactics instead of actual people. And when more than 5% of all of chat GT's messages are healthcare related, I mean, we're still early on people using ChatGPT. 800 million users, one in four submits a healthcare related prompt every week. And there's a problem with that is most of us are tracking the wrong prompts. And the reason why I feel comfortable saying that to you is because it was me just two or three months ago. You know, when I type in that I'm looking, I'm gonna share my, I'll share something here. Um, I'm looking for a medical, uh, I'm looking for a concierge medicine provider in Philly with a focus on sports medicine. And you can see Claude is like, Ooh, Claude knows things about me giving your marathons and the data you track. This is what I came up with [00:10:00] ultimately Claude, for on this answer on this day. This is why you wanna run searches the same search over and over again every week or so. We'll talk about that later. But like it actually did not have the doctor that ended up that I ended up picking, which was, in my opinion, the complete right one for me. It didn't even have it in its consideration set yet. And I think the biggest thing about prompts and the reason why we track the wrong prompts is I think more people not, I think I've seen the research that shows more people are putting brands in their prompts already. Well, how did you even know that brand in the first place? Friends, family, billboards, tv, outdoor, uh, content that you read that was high quality. That's what makes you know, a brand. So what I think is really important when you run these prompts, folks, is to keep an eye on what things are saying about your brand that you don't own, and then only you can decide, are you okay with that? Or do you wanna actively decrease your unowned answers about your brand? I know I would. So if you're [00:11:00] tracking basic prompts that sound more like search, like you really gotta kind of dial in your teams and go, Hey, are we tracking prompts that sound more like search? Or are we tracking prompts that sound more like what actual humans do? I'm choosing between Penn Medicine and Jefferson Health, and I think that's the big part. Penn Medicine, Jefferson Health, and I'm a runner, right? So I built a tool that is open source. So this one you can go and copy from my GitHub. And it basically allows you to compare brands against each other for a, uh, for a prompt. And what it does is it goes out and researches the brands, and then it gives you choices on like, well, what matters most to this person, this persona, let's just say, say it's reputation and whatever, and whatever, whatever. You click 'em and then you get a, you get a. A spider chart that shows you know, where you're strong and weak. Now, if I was Penn Medicine, my big weakness would be Access Inconvenience. I'm going, huh? Why does it think Access Inconvenience is weak? And then you can look at what was actually said, and it says, high access, but decentralized across West Philly and Center [00:12:00] City. That hurt me on Gemini. Okay. Now if that's true, then it's what it is. If you're like, but that's not necessarily true. We usually have most coverage within a two mile radius. Well then I would ask you, are you actually saying that on your website? Uh, no, we're not. That's what you gotta do. That's the new job. To rank an ai, you gotta take the things you know about your business and make sure they end up on your website. And what was crazy is chat, GPT nailed it. They said, Dr. Jeremy close is the differentiator. And then it says, this is what this means. What I actually really loved about this is, um, Dr. Jeremy Close is the one I picked. Claude never even saw him, so it's like, okay, how do I work on Claude? Am I getting spidered properly by Claude and indexed? 'cause he wasn't even mentioned. So with Jeremy Close, what I loved about this is I was like, yeah, these are the things that actually mattered to me a lot. And it said, Hey, Penn doesn't have this. So now if your Penn, let's just say last week, [00:13:00] last month, or you're currently planning on bringing on a doctor that would, that would have this experience, you gotta talk about that, right? Let's just say you have a doctor that's run 10 marathons. We should at least talk about one of our concierge doctors as a. Marathon runner. But if we don't talk about that on their profile, that person doesn't show up. At least being able to empathize with what a runner would want. Right? You see what I'm, are you starting to pick up what I'm putting down? You gotta say these things on the website, right? Or you can do it in other places as well. Um, but the website is the one that you control, but then this is the real banger, guys. It's like if you're at Jefferson. Or Penn, what do you agree and disagree with? And then what are you willing to try to put content on your website to be more to overcome? And of course, again, you gotta run these things multiple times over time, but if you kept seeing the same thing showing up, Hey, Jefferson's better than Penn because it's a smaller group. Well, if you're at Penn, you're like, well, I'm not gonna be a, I'm not gonna make my group smaller. Like I, that's not [00:14:00] an option. But could you talk about your group in a way that felt more intimate than maybe we are? If that mattered to you in your go-to market, then yes you could. Now, ah, this is one of those red pill, blue pill moments, y'all. And luckily I'm in a room full of real marketers, but because I'm in a room with a lot of freaking low quality SEO marketers, I have to remind people not to do crap like this because Jefferson could be like, we're gonna compare us to. Penn on all of our pages on our website, and we're gonna make a bunch of pages that nobody's gonna see. And, uh, but it's gonna flood the, we're gonna flood the LLMs with 80 pages. This company Search Bloom has 80 pages comparing themselves to 80 different agencies. That ain't marketing. It's bad. Marketing is what it is. It's not even marketing, if you're gonna call it marketing, at least we gotta call it bad marketing. And I wanna talk about bad marketing. You know, I think bad marketing is the result of not having a trust. KPI, I'm just gonna say it straight up. How many of us get [00:15:00] text messages from like, Ooh, we got a loan waiting for you. I get like 80 of those a freaking week. Because somewhere there's just no trust that matters. So therefore you can do whatever you want. When I looked at 18 to 34 year olds, I believe in this one report that I've linked to, um, they have disregarded advice in favor of friends and family. It's like, yo, jojo said what? This person went to school for how long, and now I can only imagine how much worse this is in an LLM world, which is why it's gonna be more important than ever to get your doctor's opinions out there on the website. Now I know that's rife with like legal and all that stuff and we'll have to figure that part out, but I would build a. If it was me, I would take what I built and I would add a legal component so that I know what my lawyers do and don't approve. And I would make that an eval as part of my interview process. So after I get my document back, I would have something flagging all the legal stuff, and then I would take those to the legal team and see like, is this still a problem? Yes or no? And then once I [00:16:00] learned. About what I have to say and what I can't say. Because it's an AI and it's real time interactive, you can run an evaluation against somebody's answers when you interview them and have the legal evaluator say, Hey, you gotta ask that question differently. Because I didn't like the way they answered it, and it would just say, and imagine on a voice call. Uh, I mean, I gave it to you guys, um, at the core, but like you could build out something that manages that extra workflow and make it easier for you to get your content approved. Anyway, I just came up with that on the fly. Um, so it might not be a great idea, but then also people are doing this. It's with social media, like you're disregarding what you heard because somebody on social media. Oh no. But it's gonna come back to be pretty important in a little bit. And the other thing that I think we need to start talking about more than ever is the compression of the funnel. You know, in earlier marketing what would happen is people would search and you would see people early in their journey. And then you got [00:17:00] cookies or whatever, and you could track them, and then you could, you could say, I'm gonna market to these people because they're at this point of the funnel. Now what I think is happening, and what I'm seeing in my own business is people are showing up so educated already know what the questions that they want answered are, et cetera, et cetera, et cetera. So now they're much further down the funnel when they get there. So you gotta change your organization to meet people where they're at more than likely. Now, the thing I've been preaching for a while is that. The biggest problem with generative engine optimization as I see it, is it is bought. By performance marketers who are used to SEO. And when you're used to SEO, you're like, man, I could track everything. So it was easy to track SEO, right? But you can clutch your pearls. I'm here to say that SEO is definitely a performance channel, and GEO is not a performance channel. And I'm sorry, but I will defend that by telling you that the math don't add up. In an SEO world, you had monthly search volume and then if you got ranked position two or three or 10 or whatever, you knew you would get a percentage of those [00:18:00] clicks. So now you knew how many visits you were gonna get to your website. And then once you get visits you're like, well, I can time. I can multiply my visits by my conversion rate and get a number of conversions, which now means I can just literally tell my boss what my ROI is gonna be on going after this back basket of keywords. None of that exists in geo, none of it. So it's so hard for you to come. So don't use an SEO frame in a geo world because you're setting yourself up and a lot of snake oils being sold to you guys. Like, oh, I got, I got a way to get back in a monthly search volume. I'm not buying it. Right? Nobody knows what personalization is doing 'cause you can't see it. Click through rates, which I'm gonna show you in a little bit, are all over the place. The next thing is the launch dates versus the training data. So we're gonna get into this later. But one of the most important things is these, you're hearing, your bosses are hearing about things. Oh, this new version of Gemini launched. What they never know is, do you know that it, that version was trained off of data That's 14 months old. [00:19:00] Alright? Which means that for some of your answers, for some of the work you've done, it has not even had a chance to show up in the answers 14 or 15 months later. We have to remind our leaders that AI visibility is not the goal. It is to be. It's just being seen. Use this framework to remind people over and over again, the job is to be seen, then believed, then chosen. So yes, it is important for me to be visible in ai, but when the model updates, when the model hasn't updated in 16 months and half my prompts don't trigger web search. All that work I did is not showing up for half my prompts. And it's so early guys that things are changing so fast. You know, many of many people woke up in November or December and were like, woo, my AI strategy is crushing it. And I'm like, no, no chat. GPT. Just over double the number of [00:20:00] words per prompt. So all people got was a longer list of competitors. So you, you might be competitor 15. So yeah, you showed up. I didn't Google all 15 of these. And for my analysts in the room, look at Dominic Pori Construction, LLC. There is no link next to his name. So for me, what I would do is I would highlight that name right click and say, search Google, guess what? Now showing up in generative AI is gonna result in organic branded search. But, but how would I attribute that back to ChatGPT? You can't dog. And then some of them are linked here, which is great, and those ones show up as attributable by chat, GPT, this is the world we live in. This is the world we live in. You know, it's like, uh, data is just disappearing everywhere and people are asking us tougher questions than ever because their frame is stuck at a time where all that data was available. The next thing is, you know [00:21:00] when you used to see underlines on Google, especially back in the Tin Blue Links era, you knew there was 10 competitors that were gonna show up. And if there was a blue link and you clicked it, you went to the website, change your frame. AI overviews when a link is underline, doesn't take you to Sun sama, it takes you to Google with the brand name Sun Sama being as another search. So you are gonna end up paying for more clicks. And keep this in mind if you're going after a competitor campaigns, it's crazy what they're doing now with those AIO reviews to juice, the number of people searching, and now your branded, I bet you your branded campaigns aren't performing as well. If you've been wondering why this could very well be it. So then when I do a search around doctors or physicians, all those links, none of them go to their websites. They all go back to Google. Where they then show your competitors' ads against that person. It's frustrating times on that. [00:22:00] The one thing I will say is traditional SEO matters for sure on near me queries, and BrightEdge did the research here. So, um, they did the research that showed that Google has really degraded near me queries in AI overviews, and when I do most of the searches, that is true, however. You might say, Hey Will, well this is an a IO review for a what appears to be a near me search or a whatever search. And I'm like, yeah, you'd be right. It seems like in Bing that is not the case. They will show a IO reviews for doctor searches near me in Bing. Okay. But in Google I checked it a bunch of times. They do not. So that means that doing traditional SEO, especially for your local and near me queries is super important. Still. I know you might be a little bit overwhelmed, so now I wanna help you to figure out who, how do you get better at figuring out who can actually help you? Right? And that's why I wanna teach you how to assess the helpers, [00:23:00] first of all. Don't listen to bro ass bros. You know this. Oh, I replaced this person for 8,000 a month with my $30 open claw. Stop it. You know, my first thought is then that means you were overpaying for talent. Why don't you change your post to say, I overpay for talent? 'cause whatever you were doing was not marketing. Because most times when you see people write stuff like that, they're doing stuff like this. And I don't think this is marketing, and that's why we gotta remind these freaking consultants that the job is not just to make me visible on ai. Sure I can play that whack-a-mole to get my balls off my ass this quarter. But I have examples, especially with ramp who did this, this, this section of their site is starting to, I think it's already lost 70% of its traffic or visibility since when I first started tracking against them. So like these things are temporary because it's [00:24:00] whack-a-mole. Do you really think that open AI or chat GPT or Claude or Gemini want this kind of low quality stuff to be how people get answers? No. Do they got a bunch of smart people there? Yes. Are they smarter than you and me? Most of them. They're gonna figure out how to knock you out, which means you're back to playing freaking whack-a-mole. I'd rather build a long-term brand, wouldn't you? All right, so now I know you can't just take that tirade to your boss. So now I wanna break down the economics of being believed and chosen using our ramp versus our Michelin example. So when you get a Michelin star, you typically get about 20% more foot traffic, two stars, 43 stars, double double your traffic. Doubling the traffic to your business probably makes you more money, right? Right. Yeah. Now the interesting thing is when you look at how much they raise their prices getting three Michelin stars lets you raise your prices on average by [00:25:00] 80%. If you raise your price, you typically make more money. That's what this is about. It is not about AI visibility. So when you got RAMP doing this to get AI visibility, who's showing up to the top six restaurants in Louisiana? Again, another huge state. Who turned their brain off and just built, oh, we're gonna do, uh, AI content around states. Who's excited for that? Zombies, lifeless, soulless zombies. It's the only people interested in this, in that content, and especially when you start to look under the hood. I have chosen to not try to be visible for this prompt, even though. I have an agency that I believe does a great job at generative engine optimization because I have made a choice. I looked at the results, look at the word top. Best top, top, best, top geo agencies, plural. Why? Because I'm showing [00:26:00] a bunch of agencies and every one of these companies lists themselves at the freaking top. And I saw this showing up for everything that I was trying to rank for. And I said, I refuse to attach my name to this low quality crap. And by refusing to do that, I have also had to sacrifice showing up this way. Luckily for me, I believe that people ask their networks, who should I talk to? Luckily for me, I believe that writing a piece of content that's a listicle that I could have done in 10 minutes is not worth me coming here, spending 10 hours on a slide deck, flying all the way here, turning around and flying home. It's a lot more expensive. It costs a lot more, but like I hope that y'all, when you're looking, will go, Hey, I'm thinking of Seer Interactive for this. Not, I'm looking for generative engine optimization. Right? And when you look at it, first page, SAGE with all their spam and the mouths first page, Sage. These are companies that I have never, ever shared a piece of their content in my life. The only company on this page who I've ever [00:27:00] seen one of their people speak or shared a piece of content was I pull rank. Nothing from anyone else. And instead of them saying, how do we get ourselves to produce better content and make ourselves a brand people search for, they're thinking, how do I build these low quality comparison pages? And that's why I wanna teach you guys how to do this yourself and how to find out what your agencies are doing as you're making these decisions. You can run this prompt or this search site, colon, and their brand name, end title, best or top, and you'll see, oh man. Scroll to the bottom. Click next. Next, next. They have produced how many pages worth of this slop, 'cause that's gonna be their strategy for you. And then ask them to open up their screen and share Google Analytics. And look, do they get referrals from places that you know in trust? Or is all the traffic to these pages just AI and Google. If all the traffic is just AI and Google, it's an indication that the content has not resonated with people anywhere. No social, no direct, [00:28:00] nothing, no email. And then ask 'em to look for a, a trend of their pipeline. Like right there on the call, Hey, uh, can you share screen and show me your Google Analytics? Oh, by the way, uh, you're in sales, right? Yeah. So I guess you have access to your pipeline. Yeah. Um, you know, go find it. I'll wait. And show me a trend of your pipeline. 'cause you just showed me your visibility of being seen in your ga. I'll see whether or not you're being cho, uh, believed. And are people signing up for your newsletter more? Yes. No. And then I can see whether or not you're being chosen. Ultimately by seeing are people filling out your lead form to get in your pipeline? A lot of these companies wanna drag you into their trap, and it's the build trap where everything is about how much content did I produce and how little time, instead of how many people have I helped to believe something about your brand? And you know, when it comes to winning an ai, what does it even really mean? Because if the way to win, like Clickup did here is to create content where you put yourself at the top and then you're so [00:29:00] lazy and you're running so fast that you leave the instructions in your posts. So now we all know I already knew it was, uh, ChatGPT, but you made it really easy. Now does that increase my trust in you or decrease my trust in your ability to help me with time boxing apps? I'm like, oh no. If your content is any indication of your product, I know all I need to know. This is why I'm begging you to start to think of having a KPI that shows your brand trust, because without it. You can spiral outta control on tactics and become the freaking Bernie Madoff of it worked until it didn't. And for a lot of these brands with ramp, it worked until it didn't, and now they're losing traffic to all those pages. Here's a test that someone ran. I think these, these tests are good. They're not doing this for anything other than testing, but they took an old piece of content, just added all the dates to be 2026. Boom. Increased traffic [00:30:00] significantly. This is that kind of red pill, blue pill moment. You know, I would rather have, I'd rather build an AI that can call a subject matter expert and interview them for 10 minutes on a topic. And then also have a human spend a half hour, an hour massaging those answers to, to make quality content, then to do it in 10 seconds with my open claw and just produce this low quality stuff. So you see, I think that we're all gonna have a moment in marketing where we're gonna have to stand up for what we believe in as marketers or just go along with the train. And like I said, I'm slow playing, maximizing my own AI visibility. And you might be thinking, man, that really sucks. That we'll almost have to say, I'm just gonna wait it out until it gets better. 'cause I'm not gonna do that stuff. But it doesn't suck for us because we've become a curator and people are starting to believe that we're a company worth listening to, and that causes them to put our brand name [00:31:00] in which. Search Bloom has actually built content comparing them to us, so it's affecting the model, which I can't control the low quality stuff. Search Bloom puts on their website to compete with me. I can make sure that I get my own stuff up and I'm trying to win in these places, right? Because eventually we're all gonna learn after you saw some of the stuff I've shown you today on who wins and how they win. I bet you a lot of you are like, I'd rather ask my friends on WhatsApp. I think that there's gonna be this trust issue that all the ais need to figure out. How to tackle, because right now the way the answers are served up,, I think there's skepticism, at least even for me, on some of the answers that I get. So what I really hope is that somehow in, in this presentation, I earned the right to get your custom instructions, and I took all the people that I listened to the most. Many of them are not, none of them are in my company except for one. And I want you to copy and paste this and put it in your custom instruction so that you get higher quality answers about geo and AI in your marketing [00:32:00] campaigns, because the generic advice is being flooded by the low quality stuff that I've been showing you. And all you gotta do is go to settings, click on personalization, and paste in the custom instruction and paste it in your custom instructions. And I've seen this work decently well. Yeah, I've seen it work decently well. I've been running some tests that I'm not able to share just yet because I gotta do a little bit more work to make sure that it's, that it's all legit. But you see like being trusted to being your, being your custom instructions. To me, that's all about being believed and chosen. And you know what? That leads to this. Look at the growth in our number of contacts. That's why my pipeline's growing because we're being believed. My direct traffic where people are just like sending links to each other in email and WhatsApp and Facebook groups and all those places I can't track went from 25,000 in Q1 to 2024 to 34,000. This is people sending stuff to other people. That's why I like direct. Referral [00:33:00] websites other we're getting more traffic from other websites. At one point, we almost tripled our traffic from Q1 of 2024 to 2025. We had a big report go out in 2025, but people linked to it from their websites and those websites were like, ah, refs and SEMrush and Ink Magazine and E-Marketer. And you can look at these small numbers. Look at E-marketer. Look how small those numbers are. 14. 14 total visits. Guess what? The visits wasn't the value. I'll be speaking at a very large conference that E-marketer is putting on in New York in a couple of months because they read my stuff and believed that I would be good to put on stage in front of a lot of very powerful marketers. You see, these trust channels are so much more efficient. My friends Slack teams, WhatsApp, search engine land, they're just more efficient. Look at how much better they outperform. Organic search and organic search is a ton of branded search. They still [00:34:00] outperform the branded search of organic search in terms of being believed, which is for me signing up for my newsletter. You read something, you heard something, and you said it's worth maybe me signing up for that newsletter. Social's almost 500% better for me. Then organic search is when it comes to getting a vi, the likelihood of getting a visitor to sign up for being, for believing in my brand. The, the thing that sucks though, is those trust channels, they're the hardest ones to track. And I'm just sorry, that is where we are, folks. That is the truth. So now it's time to figure out what are we gonna do about all that will. So let's get through these nine things quick. First thing I want you guys to do, protect your damn house. Start setting up Google alerts. Remember that old thing? Hello, darkness, my old friend. People on your company are putting evolving landscape and delve and all kinds of low quality crap on your website. You gotta have Google Alert set up to catch them, right? They are going to leave. Clickup still has. Here's the revised on their [00:35:00] website. You want to make sure that you understand across Google, uh, across Gemini. Claude Chat, GPT, when they revise content, what are their patterns? Make sure you have alerts for those. Take these steps and learn to disable your JavaScript so you can see what do the bots see when they hit my site. These sites have gotten so complex because they're visually appealing that now you've gotta make sure that you're ready for the age agentic world that is coming. It doesn't mean turn your whole site into a markdown. What it means is disable the JavaScript so that you can kind of figure out what might they see, what loads, what doesn't load and does it matter? And I have an example for you later, I know it's technical, but you do wanna have a base level understanding of how these bots work. And you really, if it was me, I'd say if anything, try to get on CloudFlare. They are very, um, progressive in how they help you to track bots. 'cause you wanna know, are they hitting my content? And try, try, try to get access to your log files and I'll leave you this [00:36:00] one thing on the, on the bot side. Claude bot is what's used to train models. You gotta make sure, remember when I said like these models haven't updated in 14 or 16 months? If you block Claude bot so it doesn't train on your content, oh man, it means that you're allowing your competitors to extend their moat because let's just say. Five days from now, you go, oh, I'm gonna unblock Claude. But today was the cutoff date. Well, you could be waiting two years before that. All that new stuff gets in the training data for the first time, and I think that now the job to be done is to become like a little inspector gadget. You're like a detective now. You need to go around your company finding great things people are doing and talking about, and then get them onto the website. I was kind of mentioning that earlier in the Dr. Close exam. You really need to educate your leaders. My friends, like I think our leaders are asking us these questions 'cause they're just like, Hey, I typed this thing in and I don't see myself showing up. Right? Why aren't we showing [00:37:00] up? And this is where you have to not be scared. I need you guys to be able to, um, and ask the right questions. Do you know the launch date of Gemini three point, whatever. I was like, oh, I think it was recent. I just read about it. I tried it. We weren't showing up. Great. So you don't know the launch date. Great. You gotta keep your stuff on point. You need to know the launch dates are the three major models so that when somebody asks you these questions why they're not showing up, you at least know the dates. That's just three dates. 'cause you just gotta know Chad ccp, t Gemini, and Claude. Okay? So they're gonna say, no, I don't know that. The next thing you have to ask 'em is, do you know the training cutoff dates for those three models? Most recent models? You need to know that, right? Because they're gonna go, uh, no, I don't know the training, what's training, what's training cutoffs? They don't know. And that's okay. Our job is to educate 'em. And then you've gotta say like, do you know if you triggered personalization or retrieval? Augmented generation? That's just two definitions. You know what personalization is? When I type in, I'm looking for an agency. Sometimes the reasoning of the models will literally show, I know you're the CEO and [00:38:00] works there, so why would I tell you about a company that you run that makes no sense. Right, right. So it will reason that out sometimes. But also with retrieval, augmented generation, it means it went out to the web to get answers. Well, if it, if the thing they asked, it didn't go out to the web, it stuck on training data, which may never have even seen the light of day yet. And I think these are the kinda ways that you make your executives smarter and bring them along with you. Alright. The next thing is interview your SMEs. You know, I wanna share with you the story of how I even got to being like, huh? Interviewing doctors, or not interviewing doctors, interviewing subject matter experts came from me looking at content on my own website two years ago, and I'll link to the Seer of sameness post that I did, and I looked at it and I said, we're ranking number three. So that's a good thing. We're a search company. We rank number three for S-E-O-R-F-P. That's a good thing, right? Wrong, because. [00:39:00] That's not the job. The job wasn't to be seen. It was also to be believed and chosen. And when I read the content that we wrote to rank for that term, I said, I wouldn't say any of this to my friend over a beer. If you asked me, Hey, Wil I'm thinking of picking an SEO company, what questions should I ask? I'd have been like, I wouldn't ask any of these questions. And I went, oh my gosh. So then I torched the whole post, got rid of a number three ranking, uh, page for S-E-O-R-F-P and said, I'm gonna redo this content based on what I believe. I lost the rankings. They came back in a couple weeks and I think we're now sometimes number one, you can do a search. Tell me if I'm just showing up. I don't know. Um, but I wrote it from the heart on what I believed as an SME. And the way I did it is I had a, I took the old post, had it ask a bunch of questions. Um, I had a developer who built this for me. This was two years ago before I could vibe code, and, um, and then I walked around and just talked into my Google Doc. That's what I'm saying. You guys can do this now, but it requires you to take the red pill and be like, [00:40:00] this content isn't good. It might be working for me, but it's not good. So I am going to interview people. And they can interview on their own time. 'cause you're just gonna send them Google Docs. Here's a question, here's a question, here's a question I will tell you. You do eventually want to get to the phone call level because then the tools allow them to change answers and ask different questions based on what's happening on the call. But anyway, we're running outta time. I don't want your blog to become the land of misfit toys for so long. Content strategy and content marketing kind of became like the, put this at the bottom of the page because I really don't wanna put it somewhere. Or Ah, do I have to link all this crap or do I have to say these words? It's like, I think for so long our blogs have become this land of misfit toys that I really wanna change that perception. And I think if you interview SMEs, you increase the likelihood of the content being stuff that people will really want to read, people will really want to talk about. And usually all my research is showing that when we get these SMEs talking. Man, they give us so much more information than we initially asked for that [00:41:00] we can use in so many places. Social, uh, we use some of it, um, in newsletters and oftentimes my clients end up learning more about their own businesses 'cause they didn't know these things because we weren't including the subject matter experts. So what I want you to do is go find an old piece of content on your website, drop it into your favorite LLM of choice. Read it first, right? And then ask it. Hey, what has changed from today to when this thing was posted? Get a bunch of questions. Run this prompt. Get some questions. Put it in a doc and tell us. Tell one of your docs. Put it in a doc, and tell one of your docs or one of your researchers, or biologists or chemist. Can you just like next time, um, on your drive home, just hit the record button and just talk through. Whatever we're allowed to use due to hipaa, just whatever your, whatever your thoughts are to these questions, you're gonna get so much high quality content out of them, and you're not gonna feel like you're doing the RAMP AI slop version. You're gonna bring it up a level. [00:42:00] You're also gonna have to review those prompts. Guys, you know, I run the prompt. Tell me about Seer Interactive at least once a week, and I've been doing it for three years. All of you've gotta find your own one or two or three prompts that you run across the three major models. What it helps you to do is start to understand their how their training changes, answers, and you start to be like, huh, they always seem to look for this. They always seem to look for that. You want to develop an intuition, and I give my clients generative engine optimization reporting all the time. So we give 'em weekly reports. Those reports are distilled down to give you reports, things you can look at, graphs. You all should literally be doing some of these manually, both incognito and with personalization because you'll start to develop an intuition. So now when you're in a room, you can be like, um, I believe this because you're seeing it week to week to week changing. Alright, gonna get off my high horse on that. Remember, your prompt should probably at least start to include some comparison prompts of brands against other brands you know, who you compete with. [00:43:00] Start putting those brands into your prompts. And I know from probably most of you, you don't have those in. Because we learned this the hard way because we have UX professionals, our clients will sometimes hire us to watch their people to structure tests and watch their people try to solve problems using, it used to be search. Now it's LLMs. That's how we learned, oh my God. People put in a ton of brands and their prompts and we don't track brands. That became a competitive advantage for our clients because we were early in this concept of, oh my God, people don't search the way that we think they search. Because we had to have UX professionals asking questions. We had to have progressive clients who were willing to say, I will pay real money to have a real human sit and ask and watch another human do these things and document those things. And now there's a tool that we are using, um, that's brought down the cost pretty significantly because it's like a, it's an AI is watching the work, uh, being done as people are, are typing in what they want outta the ai. It's been [00:44:00] really, really helpful. The other thing, if you have sales calls or if you've got paid search that you're running sales calls, what I love about them is you get to hear from the customer and their own voice. So you get to hear how they describe their problem. Same thing for paid keywords, right? Black, O-B-G-Y-N came from me looking at paid keywords and saying, why is this word black underperforming? Oh my God. People are starting to search for their doctors by their race, especially after the George Floyd stuff, especially also in mental health. So now take your prompts, ask your geo agencies, can I get a list of all my prompts? And then you go take your sales calls. If they're not progressive enough to think this way, you go take your sales calls and say, Hey, what's missing? And then you can come back to them and go, here's 50 more prompts. I think you should be tracking and you can do the same thing with your paid keywords. And I think. The reason why I think this is so important is because the brands are gonna be starting to be put in LLMs more and more and more as people start realizing that these answers are easily manipulated. Harrit is one of the people I told [00:45:00] you to follow earlier. The thing that he did is he noticed that, um, the AI. It was consistently saying a thing about one of his client's brands that actually wasn't that true. So then here's what he did. He said, we're gonna make, he said something about how their implementation was slow. He said, well, let's go build an implementation page. He published a guide to implementing the business or implementing the software. They created a video with the founder and now. They took all that, made a page and now they have overcome those kind of negatives and now they have the content that is like showing up when people ask these questions in their own voice. And that's another thing I wanted to remind you of is AI's gonna talk about your brand no matter what. So either you start writing stuff about your brand so you are on the field and giving it a shot to try to influence it, or you can just let them predict it and make it up with no grounding at all. Or let your competitors even worse be the ones to dictate things about your brand. So now the next thing I want you to do is [00:46:00] review your reviews. And Amanda tda said it best that yeah, third party content's gonna start mattering a lot. Look at these reviews for one of the doctors. Uh, for, for one of the hospital systems, always excellent and thorough. Makes every visit amazing. None of my issues were dismissed. I already referred her to three people. Man, those are some really good reviews, right? And like this is like kinda like a human condition thing. We know that people will find reasons to not buy from people they don't like. That's gonna be a big one. That affects social a ton for medical professionals. And I'll show you an example in a bit. But let me be real. The way that I came up with this one is, again, my own business. A a prospective client left us and a competitor in a dock, and I knew what the competitor's name was. So I took the competitor's name and said, if you're picking between Seer Interactive and this competitor, which one did you pick and why? And we started tracking some of the things that came outta that, [00:47:00] and we started seeing. Account manager turnover, account manager turnover, and asur, I'm like, is our, so what I didn't know as CEO is I said, is our account manager turnover that bad? First? It notice smart as these LLMs are, they're not saying here's the average account manager turnover. They don't tell you that. They just say hi because they saw the word high next to account manager turnover somewhere. So now it's my job to say, here's the industry average and here's our number publicly. So then now at least I'm. Fighting against this one review that was done that is showing up now for 38% of my branded prompts. For the first time in my life, I have actually started asking clients to give us reviews on a few of these sites that are showing up a lot because before I didn't care. I'm like, one bad review. Any human worth, a dam that is referred to us especially is gonna know like there's no way, like one bad review in 24 years. I'm okay with that. But an LLM ain't that smart. So it's like they have high account manager turnover. So, you know, these LLMs aren't [00:48:00] smart enough to know that the one bad review on Clutch, which was in 2015, we've, in 2010, I think it was, um, it's been 15 years. We haven't had any reviews. I don't know. Maybe we're doing okay, but it didn't pick that up and it was just like these reviews, a high turnover. So Nikia Barrow, the sad part about MedStar Health is she was the doctor that had all the great reviews, not 4.9 out of five, three hundred and ninety seven reviews. Remember how I told you how to stop JavaScript or how to disable it? You hit F 12 and then you hit control shift and P, and then you type in disable JavaScript and you hit enter. And when you do that on your own websites, you would've found that for them. When I disabled JavaScript and I clicked on the show more link to show 82 more positive reviews, the JavaScript broke and sent me back to the homepage, which means only three of [00:49:00] the 82 reviews are even indexed. By many of these LLMs. So one, this is why the technical audit is so important because you want to make sure that technically the good stuff is getting in there. Now, the dirty part of me says if you've got a bunch of bad reviews on your website, you might want to put JavaScript on those so they're not indexed. But that's the bad part of me, and we're not gonna talk about the bad part. All right. The last thing I've got for you is reminding you to review Reddit as we start wrapping this up. When I had done that time boxing app example, and Clickup showed up with their listicle, one of the things that I did is I said, huh. People are looking at time boxing app in Reddit, so I know people are using Reddit. So I took all the Reddit links that I could find and I put them into Claude and I said, how often is Clickup mentioned when people are making recommendations to other people? Guess what? That answer was Zero. But all these other brands were showing up a bunch and Clickup didn't even put them on the list. [00:50:00] How good is that list now? Exactly. So even for OBGYNs and doctors, we're seeing Reddit showing up. In this instance, there's a Reddit forum where people are talking about OBGYNs and they talked a lot about the ones that they missed that had shut down at some point. And yes, you saw it. It is Philly. Bitches With Taste is actually the, uh, it is the Reddit. Uh, anybody here from Philly you will know. Um, that's very one brand for us. Uh, but if you are a local O-B-G-Y-N. Philly Bitches with Taste are impacting your AI visibility, so you gotta know what's being said over there about yourself at a minimum. I know it's hard to engage on Reddit for a lot of brands. At the minimum you should know because it does affect what shows up in the AI answers. You gotta review those KPIs. You gotta make sure that you understand with what's going on with direct traffic, because it's the catchall now, it's teams and WhatsApp, and Slack, and Zoom, and all these places just showing up as direct, which means it's gonna be harder for you to make these [00:51:00] analyses than ever. And now I'm gonna say, you know, you gotta start to turn your doctors into influencers. You know, if I came back 10 years from now and I got to see you guys again, I would hope that happened for some of you. And why? Because even though this bro sub ham is uh, saying leave everything to AI and just build distribution, slow down sub him, bro. But there are parts of this bro that make a little sense. You know, it's like distribution does matter. And I've been telling my clients for a while that you wanna get ahead of this whole game. Start building internal influencers, because that's where the trust is. That's where believed and chosen is, right. And they always go, but what if they leave Will? And I put all this time and money into it and I was talking to my buddy, John Shahada and I've linked out to some of his stuff and he is like, what if they don't join you in the first place? And I was like, Ooh, John. Great point. Right? Like, and I wanna share with you an example of a person for whom this is the case. So Natalie Crane is, um. An [00:52:00] alumni of mine, it is his wife and I went to go visit them in Portland a while ago. And, uh, he showed me her Instagram and I was like, what? And I was like, this is great. And she's like, well, my hospital doesn't love that. I put all these things on Instagram. And if you're thinking well, and look at, look like for Natalie Crane, the number two result for her is Instagram because she's really warmed this thing up to the point where the social has been used enough to where it shows her work on ig. But she also does these really notice she's not like in a white lab coat sitting somewhere in a conference room. You should, this is how you should pick your surgeons and I love my job. No, she's in her robe coming outta the shower in one where she's like in a robe and she's just like, yo, I'm thinking about finding a surgeon and how to do it on social media. She's meeting people where they are. Many of us roll our eyes when we say people are picking doctors on social media. She's leaning in and being like, if you're gonna do it, here's the way to do it. [00:53:00] That's why she has. A five star rating. She feels approachable. Now, what's interesting is when I sat in their living room and I talked to Natalie, the craziest thing that she said to me was, one, people feel more confident when I'm doing surgery and it's facial surgery. She has a lot of people that have cancer where it like affects the bones and the face. So when I'm doing facial surgery, I'm showing before and after pictures, and the hospital had to get used to that. They had to make a decision that that was gonna be okay for her and she blurs out faces and all that. But what it does is it makes people feel more confident picking her because they can see what she's done for other people and feel confident. And she feels like a freaking human and a person in these videos. So I think this is gonna be the future. I really do. And she's showing up in the video pack. These videos are training AI models. You don't think they're not using Instagram in these videos to train their meta ai. You better believe it, which means she would be much more likely to show up because her name is gonna be mentioned [00:54:00] more around facial plastic surgery. Right? Especially in the region that she's in. I think this is part of the future guys, and the data is showing that people are already choosing doctors based on their social media presence, or at least it's influencing them in who they choose when they're selecting new providers. So start to experiment, y'all. I optimized my footer. I noticed when you type in, tell me about, tell me about, tell me about year after year after year, you start to see there's a pattern. I noticed a pattern and it always was like, sir is remote first. And I'm like, I mean, that's cool and all, but people don't pick me because I'm remote first. And I went, oh my God, it's in the footer of every page of my site in that constant repetition, LL m's, like I guess it's important. So we're just gonna make sure we tell everybody you're Philly first. We changed it to 130 clients, so 97% retention rate, and that showed up within 36 hours on chat. GPT. Look at all those sources for us. The awards that we've won are coming from the award sites and from our site. This is what [00:55:00] I want to try to have the most control over, right? This is what it looks like when you're really actively trying to control what is said about your brand. So what I'm gonna leave you all with also is this, uh, template here that enables you to basically look at the. The content you have published that performs well on social, and I found one three new KPIs for AI search. Great. I know that people like it because it's performing well on organic social right, but it's not doing so well on generative ai. What I will tell you is I find that when you add statistics to pages, they tend to do a lot better on generative ai. It's looking more for facts and things like that to corroborate sources and opinions. So that's one of the things I would do to try to improve my own content in this instance, that one. And then I also realized this thing is ranking for stuff as well. So why wouldn't I try to be more aggressive in how I do get words on the page or write a new page that's built for optimizing around, um, search. So my friends just get working. Action [00:56:00] will produce information. You can't read your way into learning how to do this. You just gotta start to do things and I think it'll provide information. And I hope that I built enough trust with all of you today. I really do that you go. I'm gonna include that guy in my custom instructions because I feel like I want his takes and opinions and I feel that they're honest and that is how I hope to continue to grow my business. And then maybe one of these days I will start showing up in LLM answers a little bit more. But I think that showing up in people's hearts and minds is gonna be more important. Thanks so much.