No Hacks: Web Strategy for the AI Age

226: Most Brands Are Chasing AI Visibility Backwards with Alisa Scharf, Chief AI Officer at Seer Interactive

• Slobodan "Sani" Manić • Episode 226

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 47:19

This week I welcome Alisa Scharf, Chief AI Officer at Seer Interactive, to the podcast, to ask the question she fires back at every client who walks in wanting to "win at AI visibility." Her answer flips the whole project: fix what the models get wrong about you before you chase the category terms. We got into her research on why lower-authority websites earn more citations, why LLMs recommend a brand only 2.3% of the time, and the gap nobody is tooling for: getting an agent to actually use your website, not find it.

About the Guest

Alisa Scharf is Chief AI Officer at Seer Interactive, where she runs the AI practice across the agency's client accounts. Her team's research spans hundreds of thousands of pages and tens of thousands of prompts, and she argues that citations are a leading indicator, not a business outcome.

Chapters

  • 00:00 - The first question Alisa asks a new client
  • 04:08 - Brand accuracy: what models get wrong about you
  • 06:57 - Defense wins championships
  • 09:54 - The brand accuracy audit
  • 13:12 - Why lower-authority websites get cited more
  • 15:57 - Citations are page two of Google
  • 19:37 - LLMs recommend a brand 2.3% of the time
  • 21:49 - The agentic browsing tooling gap
  • 28:13 - Losing 30-80% of organic traffic
  • 37:31 - How a 15-year-old brand catches up
  • 40:24 - What we will get wrong in 12 months
  • 43:18 - Where to find Alisa

Key Takeaways

  1. Defense before offense. Pick five factual prompts about your own company, founding, location, what you sell, who you compete with, and run them across ChatGPT, Claude, and Gemini. Fix what the models get wrong before you spend a dollar chasing category terms.
  2. Citations are a leading indicator, not a result. They swing by month and by model. Real success shows up in direct traffic, branded search, and brand recognition, none of which sit neatly on a dashboard.
  3. Visibility is not readiness. Getting cited and getting an agent to actually buy, book, or provision on a customer's behalf are two different problems. Most providers sell the first and call it the second.

Notable Quotes

"You can flip an old house and turn it into a really impressive place to live. You can't flip an old house and turn it into a skyscraper."

"Everything we just spent the last 10, 15, 20 years learning is now doing you a disservice, because you really have to turn to a fresh page and say, where is my audience?"


Resources

Connect with Alisa Scharf

No Hacks is a publication about the agentic web. Articles, a weekly podcast, and a newsletter for SEO, CRO, and web professionals who want to stay visible, trusted, and findable as agents take over. Hosted by Slobodan "Sani" Manic.

Subscribe at https://nohacks.co/subscribe

[00:00:00] Alisa: So I think executives have to be comfortable with this idea that everything we just spent the last 10, 15, 20 years learning is now doing you a disservice 'cause you really have to be able to turn to a fresh page and say, "Okay, where is my audience?

[00:00:15] Alisa: What are they doing? What are my hypotheses? How am I gonna measure this?"

[00:00:19] There's a moment in this conversation where Alisa Scharf, chief AI Officer at Seer Interactive names the first question she asks every new client who comes in asking how to win at AI visibility, the question is not what you expect, and it changes what winning even looks like. Welcome back to No Hacks.

[00:00:38] My guest, Alisa runs AI practice across every SEER Client account. Her research at scale 15,000 prompts and more than a half a million pages, found that websites earning the most Chad GPD citations are the ones which, the lower domain authority, which is the opposite of what every mainstream tool has been telling you for a decade.

[00:00:59] And she's been publicly downgrading, sedations as a metric in real time on LinkedIn, calling them in her own words. Akin to page two, Google rankings, A leading indicator, not a business outcome. A few weeks ago I had Will Reynolds on the show. Alisa's, CEO is here for the conversation about why AI visibility is a vanity metric.

[00:01:18] That was the vision level. Take. Alisa is the practitioner one layer down the person walking through what the foundation actually looks like with the data, showing why most brands are chasing the wrong target. We talked about defensive versus offensive AI visibility work, what the brand accuracy audit actually is.

[00:01:37] Why LMS recommend a brand only. Out 2.3% of the time, the catastrophic state of agentic browsing tooling, the super base pattern of agents picking what gets provisioned on a user's behalf and what marketing teams need to restructure if they want to keep up. Here's Lisa Sharp. Enjoy the episode. 

[00:02:05] Slobodan: welcome to No Hacks. It is a huge pleasure to have you on the podcast.

[00:02:09] Alisa: Likewise. Thank you for having me

[00:02:11] Slobodan: So let me ask you about AI visibility. That seems to be the hottest thing online now for the last six months, for maybe a year. When a new client comes to Seer they ask you about AI visibility, how to win at AI visibility, because that seems to be the popular question, do you ask them back?

[00:02:28] Slobodan: What's the first question you fire back?

[00:02:31] Alisa: Yeah. Yeah, it's a great question. I mean, the first thing I wanna know is what does that mean to you? What does that mean to you, and who is asking? Because often, I mean, we- we've got a bit of a luxury right now in this industry where in the past, I don't know, decade or two decades, people get SEO, they get the value, and the C-suite's like, "I don't need to hear about it.

[00:02:51] Alisa: Just do your thing, make your money." And now AI is just so top of mind for everybody, so we've got this resurgence of CEOs, you know, going on ChatGPT and doing the what's the best X for Y prompt. So they are, are placing that pressure on their team. So often people come to us and it's like, "Well, we want visibility in these results."

[00:03:13] Alisa: It's like, all right, well, where is that gonna show up on a scoreboard? Because you're not gonna get a ton of traffic. The direct traffic you are going to get, it might convert pretty well, but it might convert well because it's not actually a great qualified lead, and it might waste your salespeople's time depending on, on what industry you're in.

[00:03:32] Alisa: So I think really trying to understand what is the culture that we are selling into. Is it a data-driven culture? Is it a culture that is happy to test and learn and understands that there is not a playbook where it's like whoever can write the biggest check is gonna get the playbook and win. It's like we're all finding this out together, and we try to posture ourselves as such as, "Hey, we're testing, we're learning.

[00:03:58] Alisa: We're gonna try to keep you, you know, moving with us." But at the end of the day, this is still a very experimental channel, and it's really hard to say, "Yes, here's the, the path to win."

[00:04:08] Slobodan: love the, the, the fact when a client asks, "I want to buy this from you," you ask them why, which is not the most common way to sell things, but it's the only correct way, because usually you're not sure if they know what they're, what they're after. And maybe it's not just those best X for Y or whatever else.

[00:04:28] Alisa: Mm-hmm.

[00:04:29] Slobodan: Now you posted about this a few months ago that Will, Will Reynolds, CEO of Seer Interactive, he was on the podcast a few episodes ago. Go and listen to that, please, because it was a great episode. He does the tell me about brand prompts. He's been running those for Seer for, like, three years now, two and a half years now, a

[00:04:46] Alisa: Yeah

[00:04:46] Slobodan: long time. And when you run this on a, on a huge company, a Fortune 500 client versus a startup that doesn't have as much presence online, what are the failure modes? What's the difference in those failure modes that you notice?

[00:05:00] Alisa: Yeah, that's a really good question. So, I mean, I think a, an enterprise or a brand that has a lot of name recognition is gonna have a ton-- most likely is gonna have a ton of, of traction in the training data of the model. So there's gonna be a pretty good idea of what that model is and who they are.

[00:05:20] Alisa: Whereas with a startup or maybe a smaller business like Seer, we find that there is a, there's a, a belief of who Seer is and what Seer represents, some of which is like, "Wow, that is spot on. They are really nailing this." Other things are like, "Eh, this isn't technically true." You know? Like, we used to have a San Diego office, we don't anymore.

[00:05:42] Alisa: We're still heavily associated with San Diego. And then there's another category or degree of I hate to call it misinformation, but at a certain point it kind of is that. And we've been doing some analysis on the Seer brand. An example I'll give is Seer is not known in LLMs for having a proprietary data tech approach, which is kind of like an, you know, one of the criteria we think ChatGPT is looking for.

[00:06:08] Alisa: Like, you're a digital agency, you should have some digital agency tool that people like. And we're like, wait a second, we've invested literally millions of dollars in this data infrastructure. It has always been a part of our core. Have we done everything we could to tell that story both on our website and on third-party websites?

[00:06:28] Alisa: Not really. And I think that's what we're fighting against now. So you really have to wade through. If you're an enterprise brand, there may be a deeper understanding of who you are and what you're about. Even that could be outdated though, because so much of this is, you know, it's based on the training data or it's based on information that may not actually still be accurate

[00:06:49] Slobodan: Yeah, this is a, the fact that L-LM suggested here doesn't have all that stuff. It, it-- read your blog. It doesn't really

[00:06:57] Alisa: Great

[00:06:57] Slobodan: sense. You, you sit on so much data. Something that Will also mentioned, you don't publish as many blog posts, but you make them really count and really matter. I think that's

[00:07:06] Alisa: Mm-hmm.

[00:07:07] Slobodan: this LLM era than just pushing out content that nobody really asked for.

[00:07:12] Slobodan: Now you have I think you call defensive positioning with, with AI visibility. And I'll just say,

[00:07:17] Alisa: Mm-hmm.

[00:07:18] Slobodan: you earlier, offense wins games, defense wins championships, always and forever. All those best X for Y or just best X prompts and trying to rank for that by writing shallow articles or even not shallow, but topical articles, that's fine. But

[00:07:33] Alisa: Mm-hmm

[00:07:33] Slobodan: doesn't know who you are, doesn't know your brand, doesn't know your entity on-- ident-identity online, I, I--

[00:07:39] Alisa: Mm-hmm.

[00:07:40] Slobodan: than just trying to build on sand essentially, right?

[00:07:44] Alisa: Yeah. Yeah, absolutely. And I mean, this was really a trial and error thing that we figured out. And w- we started selling... I, I tell this story often. We started selling geo services fairly early, and we had the market, we had our clients, we had prospects coming to us saying, "I hear this is a thing. What do I need to do?"

[00:08:02] Alisa: And I was leading a lot of that work, and I was like, "I don't, I don't know." People, and this is still true today, the people who invented this technology don't fully understand how it works. So we had that kind of, you know, insecure moment of, well, who are we to try to optimize these platforms? But because we were so early to it, and Will was so bullish on testing for the SEER site, we did start to see threads that worked.

[00:08:26] Alisa: And so that gave us some confidence. But still, what we started with in selling to clients was, we wanna rank for what is the best X for Y. So that's where we start, and it's like, this is interesting. And, you know, if we update content and make sure the content is fresh, and we've got the right schema that's telling the story of, you know, what, what exactly this business is and what their services are, seems like you can get kind of far.

[00:08:51] Alisa: But we were hitting this kind of invisible wall. When we started to take a step back and kind of reverse engineer what the models thought the best answer might be, we often found, just like the situation with SEER, it's like, wait a second. The model doesn't understand that you're a payroll provider that actually does have a great collection of integrations, but you don't talk about it very well, or it's on a part of your site that's written in JavaScript, and so it's not especially indexable or crawlable.

[00:09:23] Alisa: And it's like, wait, if we can change the opinion of the model and make it so that it understands what is objectively true about the brand... And by the way, it's not objectively true that they're the best X for Y, right? So, like, you, you can say, "Hey, a lot of people like the service," but let's try to stick to the facts, and that seemed to be levers that we could just get more of a, a hand-hold on and start to see better progress than if we're, you know, swinging for a home run every time.

[00:09:54] Slobodan: I love that. It's really about, is this a website where I'm getting my answer from because there's an answer there? Or is this-- when I'm in this bucket, is this brand that always has it and the LLM knows, for example, to pull SEER when, when talking about AI optimization, AI visibility. I think it's super important that this defensive work is-- boring work.

[00:10:15] Slobodan: It doesn't give you immediate results, so I get why it's not what people reach for first, but to be done. Now,

[00:10:22] Alisa: Mm-hmm.

[00:10:22] Slobodan: know that defensive work is done on a client?

[00:10:26] Alisa: Yeah. So what we've been doing is creating these brand accuracy audits, and this is fairly new. We've got one complete for SEER. We've started to roll it out to a few clients who are really into kind of alpha testing what we're putting together. But the idea is you come up with a list of criteria, objective criteria.

[00:10:45] Alisa: Again, it can't be anything that's like, "We wanna rank for best X for Y." It's gotta be when were you founded? Where are you based? What do you sell? Who do you compete against? And you take that list of queries and see for each model, what is it consistently getting right? What is it consistently getting wrong?

[00:11:04] Alisa: And for the areas it's consistently getting something wrong, how do you dig in there? And what I've said from the beginning is we're always looking for these threads to pull on, right? 'Cause we don't exactly know how these systems work or how to approach it, but you can get a, a kind of a spidey sense from pattern recognition, from doing the searches over and over, where you're like, "Wait a second, this random website keeps getting quoted.

[00:11:27] Alisa: I've never seen it before. Maybe there's something to this. Can we go to this website? If it's got outdated information about the brand, can we update that information?" Or even better, what I love to see is when there's something on our clients' websites that's being misinterpreted, 'cause then it's like, man, you've got full control, right?

[00:11:46] Alisa: And yeah, maybe you have to argue or there's some internal politics, but that is way easier than having to change content that's long been published and you don't have a, a resource for. So you go through that with whatever models matter to your audience, and I think SparkToro is a great tool for that, to be able to identify, you know, yeah, my audience is on ChatGPT, they're on Claude, they're on Gemini.

[00:12:09] Alisa: You go through this brand accuracy work and try to keep moving the needle and hope that you can see a correlation between the areas of brand accuracy with the areas of non-brand visibility you're trying to get, and then you can take something back to the C-suite, to your clients. It's like, "Hey, there's something here.

[00:12:28] Alisa: How much more can we invest in here if it's working?"

[00:12:32] Slobodan: Yeah one of the things I, I love doing for clients is really forcing them to write-- Forcing? Not, not forcing. Telling them to write a

[00:12:39] Alisa: Mm-hmm.

[00:12:39] Slobodan: identity of their organization.

[00:12:41] Alisa: Yeah

[00:12:42] Slobodan: can you just pick organization schema and every field that's relevant, just write down exactly what should be the truth online, and

[00:12:50] Alisa: Yeah.

[00:12:50] Slobodan: thing that you said, the audit everything.

[00:12:52] Slobodan: Where does the LLM get it wrong? Why? Where do they get the information from? What do they hallucinate? then you can start fixing. And also my favorite thing is version control and that canonical identity, which why do we have version control on README files and not on identity of an

[00:13:07] Alisa: Yeah

[00:13:08] Slobodan: It, it just-- I, I don't have a good explanation for this, but I hope it changes.

[00:13:12] Slobodan: But yeah one thing Sear does is really those research pieces where you do, you publish these long articles with proprietary data. One thing that you found is that DA, domain authority, which arbitrary metric, but a metric that has been around for a long time and people know it, Websites that have a lower domain authority 20 to 40 domain authority, better AI visibility than the ones that have it 80 to 100.

[00:13:40] Alisa: Mm-hmm.

[00:13:40] Slobodan: Or, or almost as well, not better.

[00:13:43] Alisa: Mm-hmm. Yeah, it's a really good question, and I mean, I think it is a symptom of how volatile the citation space is. And the one-- the, the words of caution I always use is people like to talk about citations as a metric because it makes sense. People kind of understand what, where you're coming from with a, a citation.

[00:14:08] Alisa: But it's a really difficult metric to kind of hang your hat on because it's gonna change so rapidly, and model to model it's very different. So what we found in one study is exactly that. It was more niche websites. It was these very specific pieces of content that were influencing brand visibility, and it felt a lot like early days of the internet because you look at these websites and it's like, well, I would never go here as a buyer of this software or a buyer of this product.

[00:14:39] Alisa: But in the training data, you know, in, in some rule within the post-training data, something was indicating to the model that this is valuable. Our theory or, or my theory this whole time has been these models are going to have to really clean up the citations because sooner or later, enough people are gonna realize where information is coming from, and it's gonna cast a huge cloud of doubt over the viability of the responses that are being generated, right?

[00:15:08] Alisa: So if you do see it's like, hey, there's this... and domain authority aside, it's like, hey, there's this hugely influential website in my industry. You know, we're getting mentioned in in Ahrefs, and we're getting mentioned in Search Engine Land, but that's not being cited. What's being cited is some random like, you know, agencyspotter.io, and I don't know any of our customers went there to decide to choose us as their agency.

[00:15:34] Alisa: That has to be a thing of the past pretty quickly. And so we try not to get too bent out of shape about these, like, weird citations because by the time you figure out what's going on there, you might look the next month and it's like, all right, now we're back to Reddit or we're back to YouTube

[00:15:49] Slobodan: Yeah, Reddit is back now. Reddit is n- the number one citation source. It

[00:15:53] Alisa: Mm-hmm.

[00:15:54] Slobodan: up, it was down, now it's up again, and like

[00:15:55] Alisa: Mm-hmm.

[00:15:57] Slobodan: Wild West. It feels like early days of search, I would say. Not, not specific, but early days of Google and, and let's say Google because it's, it, it's the search engine.

[00:16:07] Alisa: Mm-hmm.

[00:16:08] Slobodan: I feel like people who were there in SEO when Google was growing, nobody in the world better qualified to solve this people who are curious enough, technical enough want to be proven wrong enough, all that stuff, because the mistakes will be made, w- doing all

[00:16:28] Alisa: Mm-hmm.

[00:16:28] Slobodan: How do you feel about that citation data in the LLMs that that's all over the place sometimes being used as a leading indicator of, of what's gonna happen?

[00:16:37] Alisa: Yeah. It's-- I think it's very frustrating because as somebody who has lived through eras of SEO where you don't fully understand the business value or you're not in the right rooms, and so you're looking at metrics that make sense to track certainly, but don't necessarily tell the right story, it reminds me a lot of when we used to report on keyword rankings, right?

[00:17:01] Alisa: And the front page of our SEO report was, "Here's all of our page one rankings and how things moved up." And it's like, okay, well, what about the, the impact to the business? And it's like, well, that's a harder story to tell because I can show you how organic search traffic is going up, but then it's a little harder to connect it back to the work that I'm doing, so I'm not totally gonna get credit for it.

[00:17:23] Alisa: So let's start here with what I, I've got more influence over. Except I would argue that citations are an even worse metric than page one visibility because they don't necessarily indicate that your brand is mentioned in that response, right? Maybe... And so we think of it as a leading indicator, where it's more akin to being on page two or page three of Google, where there's something here, like this is good.

[00:17:48] Alisa: I'm not gonna argue that. If you-- If there's a prompt that your audience is using, maybe you're not mentioned in the response, but your content is cited, that's a good place to be. But it's a good place to be 'cause there's different threads to pull on then, where it's like, okay, if I'm important enough and this content's relevant enough to be cited, how could it be improved or optimized so it tells a better story about my brand?

[00:18:12] Alisa: Or, you know, looking at the other citations, what other information is covered? How can I create more of this content? I think because AI search especially is such a difficult thing to measure with organic search, I feel like we created so much value, and I say this as, you know, I feel like a boomer SEO saying this, but we created so much value that we were only pointing to a fraction of it when you're looking at organic search traffic and revenue.

[00:18:41] Alisa: Forget all the impact we had on brands that wasn't traff- attractable. But it's like, no, we, we, we can point to something. Those are big numbers. We're doing something good. You don't have that luxury in AI search. There's nothing to point to. You might be ref- getting some referral traffic. You're probably getting some referral conversions, but I think that's probably 10 or 20% of the story, and the real story is how you're influencing your buyers, and they're coming back, and you might see it in direct traffic, you might see it in branded paid search traffic.

[00:19:12] Alisa: But that's a difficult story to tell, right? And then you've gotta talk to your branded or your paid search team about kinda sharing credit, and they're like, "No, we don't want anything to do with that. This is all our stuff."

[00:19:23] Slobodan: Yes. Yes.

[00:19:24] Alisa: Right?

[00:19:25] Slobodan: I think branding works better with organic than paid, human nature. If

[00:19:31] Alisa: Mm-hmm.

[00:19:32] Slobodan: show me this, like maybe, maybe I, I shouldn't trust you blindly. At least that,

[00:19:36] Alisa: Right

[00:19:37] Slobodan: I operate. you had another post that said that LLMs recommend a brand only 2.3% of the time. Can--

[00:19:45] Alisa: Yeah.

[00:19:45] Slobodan: seems super low.

[00:19:46] Slobodan: Can you explain,

[00:19:47] Alisa: crazy.

[00:19:47] Slobodan: that is? Yeah.

[00:19:48] Alisa: Yeah, absolutely. So there's a different-- I just love a maturity model, and I think this is, this is... I see the maturity model forming, right? There's, there's the citation where your webpage is mentioned. There's the mention where you've got your brand in the response and you're trying to figure out, maybe you're trying to pick four different banks to choose, and your bank is one of them.

[00:20:11] Alisa: That's good. But rarely is ChatGPT or Claude or any of these models specifically saying, "You should go with X." The more you have a conversation, the deeper you get, it's more likely to get to a recommendation. But a lot of people just take whatever the first response is at face value and don't even get to a clear recommendation.

[00:20:31] Alisa: And it was actually a client of ours, and we're so lucky to work with clients who push us to think even bigger than we currently think and their exec team to give them even more credit. They're like, "All right, this is great, and I get it. I'm not, like, looking for a different dashboard. Mentions are great, but what about recommendations?

[00:20:49] Alisa: 'Cause when I'm doing my searches, I am seeing some recommendations come up." And so, first of all, we're excited 'cause it's like, whoa, that's a good idea. How do we test that? Second of all, we're like, "All right, well, what is the, the benchmark for you and your brand?" And it comes back and it's, like, 2%. And we're like, "Oh no, that seems really bad."

[00:21:09] Alisa: But when you look at the rest of the competitive set and it's like, you know, two and a half percent average, it's like, okay, this is interesting. So this is what success looks like. So if we can get you to 3% or 4%, what's the impact it gonna be?

[00:21:23] Slobodan: Yeah.

[00:21:23] Alisa: Yeah.

[00:21:24] Slobodan: Just to, to, to mention the person that-- Sani Vasquez from, from your team led this

[00:21:29] Alisa: Sani Vasquez. Yes. Yes. Did the study

[00:21:32] Slobodan: th- impressive that it's this low. If you ask people, not even on the street, at a conference, at, at a GEO, SEO, whatever conference, and then you tell them it's 2.3%.

[00:21:43] Alisa: Mm-hmm.

[00:21:44] Slobodan: most of them would be shocked because I was surprised to see that this number is so low.

[00:21:48] Alisa: Yeah

[00:21:49] Slobodan: that part is really about visibility, discoverability, referral traffic, all that stuff. about agentic browsing for a second. One, one of the things that, that I'm most interested as in as a former web developer, my take is that we've been building the websites the wrong way for 20 years, and, and now it's showing

[00:22:07] Alisa: Mm-hmm.

[00:22:08] Slobodan: obvious that they're broken. You posted about no tools in this space, and there's one. We mentioned it before the recording.

[00:22:16] Alisa: Mm-hmm.

[00:22:18] Slobodan: I, I think it's a vibe-coded tool that they, they just

[00:22:21] Alisa: Yes.

[00:22:21] Slobodan: they needed something, but okay, it's better than

[00:22:23] Alisa: Mm-hmm.

[00:22:24] Slobodan: agent tr- agentic readiness tool.

[00:22:26] Alisa: Mm-hmm.

[00:22:27] Slobodan: is there such a gap, not just in tooling, but why is there such a gap in that part of, of, of the agentic experience?

[00:22:35] Slobodan: Because, yeah, discoverability, referral, all that, that-that's GEO, A-AE0, whatever you wanna call it.

[00:22:42] Alisa: Mm-hmm.

[00:22:43] Slobodan: AI goes to the website and tries to do stuff.

[00:22:46] Alisa: Mm-hmm.

[00:22:47] Slobodan: more talk about that part?

[00:22:50] Alisa: Yeah, it's a really good question, especially when you see the moves that Google is making, like just recently at the Google I/O. I mean, this universal shopping cart idea, they're trying to push more consumer applications of using agents, and people are-- public opinion about AI is so negative right now.

[00:23:07] Alisa: I do think some of these things are gonna take a little while. But to your point, this is already happening when you think about, you know, we're seeing it most in what's kind of interesting is in terminal action. So if you're a brand like Supabase or maybe MongoDB and people are going back and forth in Claude Code, you're probably a vibe coder, you don't really know right from left in this space, and Claude Code tells you, "Hey, you should use Supabase."

[00:23:31] Alisa: I'm like, "All right, go ahead, like do whatever you need to do." That's an agent running off on my behalf, setting up an account or opening up the screen so I can in-- you know, type in my credentials. I think the reason that we are so far behind in this space right now is first of all, it hit us pretty-- it was one of those suddenly, gradually then suddenly things, right?

[00:23:52] Alisa: To your point, this is not a new problem But agents were kind of just a glimmer in our eye even a year ago, right? And that's for us who's so in the weeds on all of this. And I think the market is catching up more quickly than they did with AI search because you see a lot of VC money being pushed into this space.

[00:24:12] Alisa: But almost every tool I try, I-- You know, it's like, "Oh, submit your website, we're gonna do this check." And you do the check and it's like, is this an AI search visibility report? Like, you're not telling me anything about agents. And the reason I think, we work with a great partner called Check C-H-E-Q, and I've learned a lot from our VP of analytics, John Lovett, on this, this whole subject.

[00:24:34] Alisa: Part-- One of the key issues is so many of these agents don't self-identify, so you wanna be able to look at a dashboard. And some of them do, right? ChatGPT has a few different, Claude has a few different. So to some degree you can look at your log files and see which bots are visiting your site. But often if these bots are not self-identifying, it's like, well, I don't know, the internet's a crazy place right now, right?

[00:24:58] Alisa: So there could be bots trying to be malicious, there could be bots trying to support my, you know, user story, trying to sign up for Supabase and everything in between. So I think that is the key thing to solve right now, that diagnostic work to try to identify at least current state based on the rules you have set up, based on how you're allowing bots to, to even access your site, what does that look like?

[00:25:25] Alisa: But we have this whole be seen, be believed, be chosen mentality where it doesn't just stop at visibility. And there's a lot of brand, there's a lot of brands who I think maybe never need to worry about this, right?

[00:25:37] Slobodan: Absolutely, yes. Yeah.

[00:25:38] Alisa: where it's like, yeah, it's not gonna be a thing, your audience isn't gonna be doing this.

[00:25:42] Alisa: But probably more brands than realize need to be thinking about this. Like a Macy's, I think you probably need to be thinking about a department store, that kind of thing, because especially from the shopping component, these bots are gonna need to be able to not just see what you're-- what products you offer, but what inventory looks like, what shipping looks like.

[00:26:03] Alisa: It's gonna need to know every single piece of information that a customer would need to know in order to put it in this universal shopping bag. And then my heart kind of breaks for these businesses because then it's like, well, if you do all that work and you set up your-- you, you invest in all the data infrastructure you need to, but you're not the cheapest product and there's no reason to, to buy from you for any kind of loyalty rewards or anything, then you're probably gonna get sniped right at that last minute.

[00:26:32] Slobodan: It's going to be so weird. I'm glad you mentioned Supabase. I had a chat with Will about Supabase and that's when we both actually realized that, we- we're getting Supabase every time we're building an app, and there's

[00:26:42] Alisa: Uh-huh.

[00:26:43] Slobodan: very deep dive into what they're

[00:26:46] Alisa: Ooh

[00:26:46] Slobodan: how they're doing it.

[00:26:47] Slobodan: I can, I can share details details about this as well. But basically they're what's called a Postgres development platform. That's their slogan.

[00:26:55] Alisa: Mm-hmm.

[00:26:55] Slobodan: when you think about it, for most people using Supabase and provisioning Supabase now, they don't know what Postgres is,

[00:27:02] Alisa: Yeah.

[00:27:02] Slobodan: does. So they're not talking to the human behind the agent, they're talking to the agent.

[00:27:07] Alisa: Mm-hmm.

[00:27:09] Slobodan: Was it Skills? What-- Stripe's... No, not Skill. Some- they launched something recently where you can provision a product, you can give your

[00:27:15] Alisa: Mm-hmm.

[00:27:16] Slobodan: and it can spend on stuff.

[00:27:18] Alisa: Yeah

[00:27:19] Slobodan: is going to move much faster. I had a-- my previous guest was Matt Biilmann from, from Netlify. They're going all in on being agent first. They, they, they're talking about agent experience, how to provision Netlify, how to use Netlify, and it's working for them. And of course, tech is the first sector to do this, but this is not going to stay limited to tech. I'm willing to make that bet. This is But the tooling, tooling is catastrophic.

[00:27:41] Slobodan: We, we completely agree on that. I mean, Google has that new check for agents where they check for llms.txt.

[00:27:48] Alisa: Mm-hmm.

[00:27:49] Slobodan: There was a big debate this week about whether that's an SEO

[00:27:52] Alisa: Always

[00:27:53] Slobodan: But I think SEO people missed the point there because that was not an SEO instruction. I think LLMs.txt is, is wishful thinking at this point.

[00:28:02] Slobodan: Yes,

[00:28:02] Alisa: Sure. Yeah

[00:28:03] Slobodan: it, it's a proposal. It's not even a standard.

[00:28:05] Alisa: Mm-hmm.

[00:28:06] Slobodan: we were not talking to you SEO people. This was something else, and that's a very important distinction to make.

[00:28:12] Alisa: Yes

[00:28:13] Slobodan: thing that I talked to Will about and, and you had it in a post you see all of your clients losing organic traffic, like 30 to 80%. Will shared that SEER lost 80% of organic traffic over two years, which you're an SEO agency. Like, it's not supposed to happen to you. If it happens to you, we're not--

[00:28:30] Alisa: Mm-hmm.

[00:28:30] Slobodan: all the

[00:28:31] Alisa: Mm-hmm.

[00:28:33] Slobodan: How should people deal with that business-wise, talking to their boss, talking to their client? What, what should the approach be?

[00:28:40] Alisa: Yeah. I mean, I think that there's a new reality, and I don't know how many favors we have done ourselves as SEOs to earn the trust to say, "Hey, trust me on, on this one." But if you look at the, if you look at the, the situation, a lot of brands, including us, had invested in top funnel informational content, and the primary goal should have been to help your users, but a very close number two goal is, okay, and this is gonna help search engines understand who we are, what we do, how we think.

[00:29:15] Alisa: Now that so many of those searches are met with a zero-click SERP where maybe we're cited in an AI overview, and I do think there's nuance here, right? There are, there are situations where you're searching for something, you have the option to just stay on the zero-click SERP, or it's a click that invites a query or a, a query that invites a click, and you do need to learn more.

[00:29:41] Alisa: You do need to, to continue on to a specific website. And a good example of this is we have a post about how do you set up your AI search traffic as a channel in GA4, which I think now they solve natively, but for a while

[00:29:54] Slobodan: days ago. Yeah.

[00:29:55] Alisa: Yeah. Yeah. But for a while it was like, oh, it's just an easy regex thing, but, you know, most people, a lot of people didn't know how to do it.

[00:30:02] Alisa: And so that's the kind of citation where it's like, yeah, somebody's gonna click through that to get the information. You know, eventually maybe the regex everything they need is just on the SERP, and there goes our traffic. Now, you can look at it one of two ways. You can say, "Okay, now this has zero value to me and my business, and this is all wasted effort."

[00:30:21] Alisa: Or you can say, "Well, that person who was looking, who had a problem, found their solution. Maybe in the back of their head they're gonna remember that I came from Seer, and maybe that's a future client. Maybe that's a, a future employee of ours." But that's a really difficult thing because it doesn't live on a scoreboard, right?

[00:30:40] Alisa: And the scoreboard, every metric we're used to is like, meanwhile it's going down.

[00:30:45] Slobodan: Yep.

[00:30:45] Alisa: thing for Seer and many of our clients is even if our traffic's going down, revenue might be flat or revenue and leads might even be going up. So part of the reAllisation, and this came straight from Will, part of the reAllisation we've had over the last few years is, you know, this is, this is not our core channel.

[00:31:03] Alisa: And we've always kind of known that, right? As in we ranked number one for SEO agency, and then the types of clients that hit us up is like, eh, they don't have the budget. Maturity-wise they're not quite there. They're not that enterprise fit that we're really after. But there is still some opportunity there.

[00:31:18] Alisa: I think a lot of that opportunity has shrunk and now we want attention, even if it's just a fraction. We want that attention on LinkedIn. We want that attention on social. We want that attention through live events. So I think executives have to be comfortable with this idea that everything, it's not-- doesn't feel good, but everything we just spent the last 10, 15, 20 years learning is now doing you a disservice 'cause you really have to be able to turn to a fresh page and say, "Okay, where is my audience?

[00:31:47] Alisa: What are they doing? What are my hypotheses? How am I gonna measure this?"

[00:31:51] Slobodan: It's a completely different world and almost like AI should be used more as a branding tool. AI search should be

[00:31:57] Alisa: Mm-hmm.

[00:31:58] Slobodan: tool than, than performance marketing channel almost. The search was closer to that. It was closer to PPC than it was to branding traditionally, s- organic search. They kind of go hand in hand. They, they, they, they were, they were almost, almost the same thing. So h- how to change that? Like, organization that's just focused on but organic and PPC, and that's all we do, this is how we make money. What's the mindset shift that needs to happen here?

[00:32:29] Alisa: You mean for the agency that's selling the service?

[00:32:31] Slobodan: agency or the client of an agency, whoever it is.

[00:32:34] Slobodan: Yeah.

[00:32:35] Alisa: I think you've gotta be able to, you've gotta be able to evolve, first of all. Because I think one core issue that doesn't come to mind immediately for, for most people, but is gonna be a core issue, is you're not gonna be able to attract the talent you need to win if you are doing yesterday's tactics and only focused on yesterday's channels, unless you are so extraordinary and you're in the top 1% of this stuff, and you've got this very specific lane of industry or, or service offering that you're going after.

[00:33:08] Alisa: But if you really wanna be a consultant and an advisor to a client, I think you have to kinda move where the market is moving, and it doesn't necessarily mean do it all. Like, for many, many years, we've known our clients do email marketing. They're frustrated with email marketing services, and we've always been like, "It's interesting, but we don't have that expertise, you know.

[00:33:28] Alisa: I, I don't think that's an area we wanna get into." But I think you mentioned this earlier on, this SEO and this SEO and, and paid search holistic skill set is really, really helpful with AI search because we, we've been through so much of this before, and so much of it is, you know, now you kinda have all the, you've got the full answer key from all the, the DOJ trials with, with Google.

[00:33:53] Alisa: W- so you kinda know what's going on under the hood. But for many, many years we didn't, and we had to go by, you know, what John Mueller or Matt Cutts is telling us, and that wasn't always the right direction. So you've always had to kinda read between the lines and have your own Spidey senses develop. So now I think if you've got people who are really good at that, now evolving them to think about AI search is really critical.

[00:34:17] Alisa: And if for no other reason than all of this sits on a layer of search engines, right? These things are not able to be devoid of each other, and I don't think that's going away anytime soon. I think there always is going to be a need for LLMs to validate what they believe, in quotes, with what's actually present on the internet today that is relevant to this conversation.

[00:34:41] Alisa: So there's always gonna be that role for, for search.

[00:34:44] Slobodan: I, I completely agree. And also looking at the way LLMs respond, sometimes they do the RAG thing, sometimes they don't. I think just tracking if that happens or not for a specific question i- is important for a brand because it tells you what you need to do. Do you need fresh data, topical data, or can you-- do you just need to do better to

[00:35:04] Alisa: Mm-hmm.

[00:35:05] Slobodan: data set?

[00:35:06] Slobodan: This is really interesting. You mentioned email marketing is a completely different channel. Can, can a channel silos, that still be a thing moving forward? Can, can anything be a completely separate thing in an organization, or does everyone now have to work together?

[00:35:21] Alisa: I think we, I, I, I think there is no way you can run a functional marketing team that exists solely in silos. And there's still, there's always gonna be a need for bureaucracy and internal politics when you are dealing with lots and lots and lots of people, especially smart people with opinions and different perspectives.

[00:35:43] Alisa: So I don't think there's an ideal state where it's a utopia and everybody, you know, lives in perfect harmony. But there is so much valuable information that each one of these different groups holds, not just for the purpose of, like, growth marketing, right? Where there's something-- There's insights email marketers have about your existing customers relative to your prospects.

[00:36:09] Alisa: There's so many insights search marketers have. But the other thing that I'm hearing that's really interesting is as search marketers, we've always been, you know, getting our hands dirty in spreadsheets or building little scripts or automations, and so we've adopted a lot of these AI tools really willingly.

[00:36:26] Alisa: Not everybody, but I see a lot of SEOs leaning in. And then when you do kind of open the door and see what's going on with email marketing or brand or corporate comm, and they're like, "Oh, what is this meeting recorder that's, that's present here?" Or, "Oh, what are you guys... Like you're actually using your Copilot subscription."

[00:36:44] Alisa: It's like, okay, there's a lot, you know, this, this tide can raise all ships if everybody's on the same page with visibility and alignment into what this future can really be

[00:36:55] Slobodan: Yeah, this is the I-I've never seen so many mid-career SEO people happy li-like they are

[00:37:01] Alisa: Yeah.

[00:37:01] Slobodan: because this is a toy for them. This is, this is

[00:37:03] Alisa: Why is

[00:37:04] Slobodan: to, to the youth days from 20 years ago. And now I can do this thing again, and I can do it properly with all the knowledge I have. It's,

[00:37:12] Alisa: Yes

[00:37:12] Slobodan: it's beautiful to see.

[00:37:14] Slobodan: And I've been trying to get as many SEO people on the podcast lately because I think the most important group of digital professionals is SEO people right now. If they're going to-- You are going to get us out of this and, and into the, the clear seas

[00:37:27] Alisa: Mm-hmm.

[00:37:27] Slobodan: know and will. That's what I'm

[00:37:29] Alisa: Yes.

[00:37:29] Slobodan: to say here.

[00:37:31] Alisa: Yes. Yes

[00:37:31] Slobodan: So let's talk about if a brand has been doing the non-AI approach for the last 15 years, and most have, and now they're trying to catch up and, you know, they're, they, they haven't been consistent online, they haven't been saying, telling the same story, how do they catch up? Like, how do they fix all that's been done in the past?

[00:37:49] Slobodan: Like, what's the, the, the, the playbook there?

[00:37:53] Alisa: Yeah. It's a really good question. And I think it's highly dependent on how much top-down buy-in you have and what the internal culture is you're selling into, as well as who is your audience, what is your business model? You know, if you're in publishing, all the odds are stacked against you. And so I think that's a very, very different digital strategy to deploy in 2026 versus if you have a small business, your audience is kind of emergingly adopting more of this technology, you've got more of a roadmap, and I think you have a little bit more time.

[00:38:30] Alisa: I think still starting with the brand accuracy work, if you wanna be visible in AI search, that is such an important thing to do. The other thing, and this applies to any kind of organization, and sometimes enterprise brands are most guilty of this, is positioning work. There's a lot of brands who have been like, "You know, we're a household name.

[00:38:48] Alisa: Everybody knows us. We're for everybody. We're this conglomerate." And then when we have these brand accuracy interviews and conversations and try to get, "Okay, what is your UVP? What is-- what's the core belief?" It's like, "It's just corporate jargon. We don't really have one." That's gonna be really difficult 'cause these models are trying to be matchmakers, right?

[00:39:08] Alisa: They're trying to say, "Okay, Alisa, this is specifically what you're looking for and who you are, and I'm gonna get the perfect option." And if there's one brand that actually is a pretty good option, but it's not clear it's for me, and there's another brand that's smaller, more niche, and is exactly for me, maybe I should have gone with brand one 'cause they've got better pricing, better resources or whatever, but those more niche brands are often winning.

[00:39:33] Alisa: And so doing that work, starting that conversation of, "Hey, this is gonna be something we need. Let's get the board on, on board with this. How do we start this process?" 'Cause that's the kind of thing you could spend years arguing about what your, your positioning and UVP is, especially if you've had the luxury of selling to everybody, and it's like, take as long as you need, but we're not really gonna be able to move the needle until some of this foundational stuff's done.

[00:40:00] Slobodan: I think the fact that in 2026 still need to ask who they are, like brands need to ask who they are before they can proceed, it's just, just mind-boggling. And another guest on this podcast, Jess Schultz, she's-- y- you probably know her, at least online. I think something that-- a question she asked me is, was something like "How come Mad Men were better at branding with no tools than we are in the 21st century?"

[00:40:22] Slobodan: W- it, it's

[00:40:23] Alisa: Yeah

[00:40:24] Slobodan: mind-boggling how we took the shortcuts and everything with the, the, the paid traffic and the meta ads and just made these people richer while we forgot and, and ignored the basics. It, it, it's crazy. Future for-- future-looking questions. Question, 12 months from now, when you're looking back, the listeners are looking back, what's the one thing that we're doing now that's going to be proven wrong, most likely to be proven wrong in 12 months?

[00:40:52] Alisa: I think, I don't have-- Of course, I don't have my, my crystal ball, so I don't have... I don't know exactly where this goes. But I have long been saying we are probably under-indexing on Meta and over-indexing on ChatGPT. And both, I think, are going to be important depending on who your audience is, especially with Google's announcement yesterday.

[00:41:15] Alisa: Now I'm like, "How well is this still true?" Because I I-- The way-- The path forward I see for Meta is to get really good with wearables and tap into their existing saturation into a really big audience and have that be their way in. 'Cause you're not gonna beat ChatGPT, you're not gonna beat Claude right now.

[00:41:32] Alisa: That said, if Google gets there first, they could easily win. And either way, I think three years from now, Apple probably beats them, them all when they finally decide to, to, like, really wade into this space. But my, my high-level prediction is I think we're gonna look back at this data we have, and so many of us are tracking in Copilot or tracking in Grok or tracking in all these different platforms, and we're gonna look back and be like, "Ah, we got it wrong.

[00:41:58] Alisa: We thought our audience was going over here. They actually went that way."

[00:42:01] Slobodan: The Meta one is pretty good because I wouldn't have thought of th- about that because they have a Zuckerberg problem, I think, that they need to sort out. Like, he has an

[00:42:09] Alisa: Okay.

[00:42:10] Slobodan: now project. The Metaverse was 70 billion, disappeared, and nobody knows where they went, and like all that stuff. they sort it out, they have the

[00:42:19] Alisa: Mm-hmm.

[00:42:20] Slobodan: their Ask Meta in every app they have, which

[00:42:23] Alisa: Mm-hmm.

[00:42:24] Slobodan: that, but if that was actually good and people start using it, that's a good answer.

[00:42:28] Slobodan: My answer is really, I think operating system is above an app, obviously,

[00:42:35] Alisa: Yeah

[00:42:36] Slobodan: above Anthropic. Google has sorted this out with the announcements they had. I think Apple does this in September when they have their big conference, and then we are not talking about Anthropic and ChatGPT as much.

[00:42:49] Alisa: Mm-hmm.

[00:42:51] Slobodan: LLMs, all that stuff, that only Google and Apple can do on-device LLMs. And if those are as good as the cloud models, why do we need the cloud model?

[00:43:01] Alisa: Yeah, 100%.

[00:43:03] Slobodan: it'll be interesting. So I would focus on, on those two. But what else are we getting wrong? I think ignoring, as I mentioned earlier, ignoring the fact that this AI traffic is coming to the website and trying to do something with no human in the loop. Come on, it's impossible to ignore that.

[00:43:17] Alisa: Mm-hmm.

[00:43:18] Slobodan: most important thing moving forward, in my opinion at least. Okay. can people connect with you other than seerinteractive.com website? What's the best way for people to find you online, follow your work?

[00:43:28] Alisa: Yeah, yeah. I, I try to post a lot on LinkedIn, so I'm very engaged on, on LinkedIn. You can find me there. You can also find me on Twitter, but you're gonna see a little more, little more fantasy football and, and sports conversation on, on my Twitter.

[00:43:40] Slobodan: I like that. This was an excellent interview. Thank you so much for taking the time. I enjoyed talking to you. And to everyone listening to the podcast, to this episode, please consider rating, reviewing, sharing. Go to nohacks.co/subscribe and subscribe there to the newsletter and I will talk to you next week.

[00:43:58] Alisa: Thank 

[00:43:58] Now two things from this conversation are going to stay with me for a while. First, the defensive before offensive framing, the brand accuracy audit that Alisa described for you. Pick a list of factual prompts about your own company. Run them across the models and find out where the LMS are wrong about you.

[00:44:16] That work is unglamorous and mostly skipped. The take home is small. You pick five prompts about your brand right now. Run them on chat, g PT Club Gemini. Find out what the models. Know about you, who your customers are, where you're based, who you compete with, the answer is most likely going to be surprising.

[00:44:38] That's your audit, that that's the whole thing. Do that before you spend any time trying to win the category terms. Now, second, the gap. Alisa named in agentic browsing tool that I could not agree more about. There's a real difference between AI search visibility, getting citing the answers and agent readiness, making it possible.

[00:44:57] For an agent to actually use your website, most providers are selling you the first thing and calling it the second, and the agent as buyer pattern that Alisa and I talked about with Super base where Cloud code recommends super base to a vibe coder and provisions the account on their behalf.

[00:45:14] That pattern is moving outside tech faster than the tooling is moving to catch it. So if your business is something an AI agent might buy, book, or provision on a customer's behalf, this gap matters now and not later. And if you want weekly breakdowns of where this is going, what shipped, what changed, what to do about it, no hacks newsletter is a place.

[00:45:36] So go to no hacks.co/subscribe. The link is also in the show notes. Thank you to Alisa again for the conversation, and I will talk to you next week. Oh

[00:46:14] no. Hatch,

[00:46:33] no hatch.

[00:46:38] Go back.


Head Shot

Slobodan (Sani) Manić

Host
Head Shot

Alisa Scharf

Guest

Podcasts we love

Check out these other fine podcasts recommended by us, not an algorithm.

From A to B Artwork

From A to B

Shiva Manjunath
Product for Product Management Artwork

Product for Product Management

Matt Green & Moshe Mikanovsky