Latest / Found in AI: AI Search Visibility, SEO, & GEO / Is Your Best Content Invisible to AI Search?

Transcript

Is Your Best Content Invisible to AI Search? · Found in AI: AI Search Visibility, SEO, & GEO

• 0:41
  SPEAKER_01: You know, is this an answer problem or is this a business problem?
• 0:43
  SPEAKER_01: It's really fundamentally a business problem, right?
• 0:46
  SPEAKER_01: Because that perception in the market already existed and the LMs are just parroting that perception.
• 0:52
  SPEAKER_00: What if improving your AI visibility doesn't mean creating more content?
• 0:58
  SPEAKER_00: Hey, welcome back to Found in AI.
• 1:01
  SPEAKER_00: I'm Cassie Clark, an AI Search Visibility Consultant and host of this show where we track what's actually happening in AI Search and what it means for your strategy.
• 1:12
  SPEAKER_00: Today I'm talking with Daniel, an enterprise SEO strategist at Salesforce, who has been leading the AI Search work for Informatica.
• 1:20
  SPEAKER_00: And we're getting into something I think a lot of brands need to hear.
• 1:24
  SPEAKER_00: You may already have a huge amount of useful information that could potentially help you show up in AI-generated answers.
• 1:33
  SPEAKER_00: The problem is that AI's engines can't easily access or use it.
• 1:38
  SPEAKER_00: Daniel walks me through an initiative to take the main ideas buried inside gated PDFs and turn them into accessible answer-first content without completely giving up the lead generation value of that original asset.
• 1:53
  SPEAKER_00: But we also get into something bigger.
• 1:56
  SPEAKER_00: What happens when AI engines get the wrong perception of your brand?
• 2:00
  SPEAKER_00: How old content can reinforce that perception?
• 2:04
  SPEAKER_00: Why point of view content matters, and what Dano's team has actually done to influence those answers.
• 2:10
  SPEAKER_00: Let's get into it.
• 2:32
  SPEAKER_01: So, you know, kind of came in, um, you know, ran uh a link building department at a small SaaS SEO agency for about seven years.
• 2:40
  SPEAKER_01: Um, then for three years, I worked at a company called Informatica, which is like a data management company.
• 2:45
  SPEAKER_01: So I was the SEO lead there.
• 2:48
  SPEAKER_01: So running the, you know, all the SEO strategy.
• 2:50
  SPEAKER_01: And, you know, this was the time where like, you know, AI search started to become really big.
• 2:54
  SPEAKER_01: So starting to run, you know, AI search program.
• 2:56
  SPEAKER_01: You know, we had an AI SEO tiger team with our CMO and SVPs, where we're really trying to, you know, explore the different ways and angles that you know, Informatica was showing up, you know, how we can influence AI answers, things like that.
• 3:08
  SPEAKER_01: And then uh Salesforce acquired Informatica.
• 3:12
  SPEAKER_01: So as of March, I work at Salesforce now and I'm still the SEO domain owner for Informatica.
• 3:17
  SPEAKER_00: So are you doing the AI search strategy lead for Salesforce or is there a whole team?
• 3:22
  SPEAKER_00: Like tell us.
• 3:23
  SPEAKER_01: There's a team, but um I'm still working on on the Informatica side, but still running the AI search program for Informatica under Salesforce.
• 3:30
  SPEAKER_00: Okay, gotcha.
• 3:31
  SPEAKER_00: So I've seen a couple of your LinkedIn posts.
• 3:33
  SPEAKER_00: Some of them have been like, it's a lot of SEO.
• 3:37
  SPEAKER_00: Are you still thinking that or are you thinking that it's kind of like expanded to something a little bit different?
• 3:42
  SPEAKER_01: Um, I feel like it's, you know, depending on who you talk to, maybe people say it's 7030, 6040.
• 3:47
  SPEAKER_01: I mean, look, SEO I still think is the foundation, right?
• 3:50
  SPEAKER_01: Like a lot of um, you know, between obviously, you know, the biggest AI engine now that people use is AI overviews, right?
• 3:58
  SPEAKER_01: Because it's not an opt-in thing.
• 4:00
  SPEAKER_01: You know, your mom or dad might be using that.
• 4:02
  SPEAKER_01: You know, I can't tell you how many times I've talked to my mom.
• 4:04
  SPEAKER_01: She's like, oh, well, you know, I Google said this.
• 4:07
  SPEAKER_01: I'm like, well, you mean the AI overview, right?
• 4:09
  SPEAKER_01: So she doesn't even really have the sense that she's using AI search, but you know, she is.
• 4:13
  SPEAKER_01: Um, so I would say that, you know, obviously Gemini uses the, you know, the Google algorithm in terms of, you know, how they're looking at pages, not saying it's a one-to-one thing, but you know, do it taking what I consider like a topical authority approach is still really important, you know, trying to show up in as many different places in as many different ways as possible.
• 4:33
  SPEAKER_01: It's still really important for AI search.
• 4:36
  SPEAKER_01: Um, but there's some fundamental differences, right?
• 4:38
  SPEAKER_01: I mean, you know, AI engines are not search engines.
• 4:42
  SPEAKER_01: Um, you know, there's a lot of personalization, you know, there's a lot of personal context that's put in.
• 4:46
  SPEAKER_01: Um, you know, there's a non-determinism element where, you know, I can't tell you how many times I've talked to people and they said, Oh, I, you know, I don't like the answer that's coming up.
• 4:54
  SPEAKER_01: I'm like, well, try running it, you know, 10 times in different 10 different ways on 10 different browsers and you know, see what comes up and it's going to be different every time, right?
• 5:03
  SPEAKER_01: So there's this element of like, we, you know, I think there's been a long time that people have sort of had the sense of like, how do we figure out the Google algorithm?
• 5:11
  SPEAKER_01: And I think, you know, Google has been both um, you know, has both, you know, kind of helped to um, you know, develop the SEO community, has at times been adversarial for sure, but um, you know, there's been 20 years of people figuring this out.
• 5:23
  SPEAKER_01: And I've been doing it for you know 11 of those 20 years.
• 5:26
  SPEAKER_01: So I have a pretty good sense of what to do and not to do for Google.
• 5:30
  SPEAKER_01: For AI Search, I think it's a it's a little bit of a different ballgame.
• 5:33
  SPEAKER_01: I mean, people can't even agree on an acronym, right?
• 5:36
  SPEAKER_01: Is it AEO, GEO, et cetera?
• 5:38
  SPEAKER_01: And you know, there's a lot of things that you know get thrown out there.
• 5:42
  SPEAKER_01: I think a lot of things that get regurgitated on, you know, we're talking about LinkedIn that get regurgitated points on LinkedIn where people use, you know, AI to write on LinkedIn, and which, you know, is okay as long as you have your own ideas.
• 5:54
  SPEAKER_01: But um, you know, people will say, like, oh, you know, you have to do schema and you know, this does that and this does that.
• 5:59
  SPEAKER_01: And it's like, well, you know, a lot of people have gone and tested that now, and you know, maybe it does, maybe it doesn't.
• 6:04
  SPEAKER_01: You know, there's been very mixed results on that.
• 6:06
  SPEAKER_01: And I think there's been a lot of mixed results on a lot of things that I think people early on considered assumptions because um, well, you know, they were just slightly more technical elements of SEO, right?
• 6:16
  SPEAKER_01: But that's not totally true.
• 6:18
  SPEAKER_01: And I think anyone who, you know, presents themselves as a total expert obviously doesn't know what they're talking about because this has only really been out for you know a couple years now.
• 6:27
  SPEAKER_00: Yeah, and it changes by the day, it feels like.
• 6:30
  SPEAKER_00: Like what works today did might not work tomorrow.
• 6:32
  SPEAKER_00: I've seen that with my own tests and experiments.
• 6:35
  SPEAKER_00: So, in your intake form, I want to go back to something that you said specifically.
• 6:40
  SPEAKER_00: You your last line of what would you actually like to talk about on the podcast?
• 6:44
  SPEAKER_00: You said we shouldn't just make a ton of content for the sake of it, but rather we should make the content we already have easier to access and use.
• 6:53
  SPEAKER_00: You want to expand on that for a second?
• 6:55
  SPEAKER_01: Absolutely.
• 6:56
  SPEAKER_01: So, one of our big initiatives at um Informatica under Salesforce now is um kind of like an ungating strategy where um you know we're trying to create an answer first layer.
• 7:07
  SPEAKER_01: So I think for a lot of SaaS companies, and you know, I've been talking about this a lot, is um, you know, they have so much content that lives behind PDFs for gated.
• 7:16
  SPEAKER_01: Um, that strategy made sense when you know it was just SEO and paid search, where you know, you would kind of publish your say your headterm articles, you know, kind of use kind of like a HubSpot 2011 inbound marketing strategy, right?
• 7:29
  SPEAKER_01: And then you would, you know, use paid search to draw, you know, eyeballs to these gated assets, you know, once you kind of warm them up with um the SEO content and retarget them.
• 7:38
  SPEAKER_01: That makes sense.
• 7:39
  SPEAKER_01: In an AI search first world, that funnel's broken, right?
• 7:43
  SPEAKER_01: So now you need to make sure that the information um from these PDFs, at least the main points, are visible on LLMs because the way people search, especially now, I've worked in um, you know, enterprise software and SaaS, you know, for my whole career, especially for people who are very tech forward who are using um, you know, looking for recommendations around software, you know, looking for pros and cons.
• 8:05
  SPEAKER_01: The more that's behind the gate, well, you know, an LLM is not going to not answer the question and say, I don't know.
• 8:11
  SPEAKER_01: It's gonna either make something up, it's gonna draw from competitors, you know.
• 8:15
  SPEAKER_01: Um, so you need to make sure that your information is in the LLM.
• 8:18
  SPEAKER_01: So what we've been kind of doing is running what I like to call like a hybrid gating process where we'll we'll more or less take the PDF, you know, we'll work with um, you know, one of our writers or product marketers, and then we'll sort of take, you know, extract the main points of this PDF and create it into an article format that, you know, is of course, you know, exists as a search asset, you know, for topical authority, it can rank, but also um, you know, preserves a gate, which is you know important.
• 8:44
  SPEAKER_01: Like we still want lead capture, right?
• 8:47
  SPEAKER_01: But we we want to ungate the main idea.
• 8:50
  SPEAKER_01: So, you know, having us, you know, essentially like an executive summary version of what the gated asset would be.
• 8:55
  SPEAKER_01: And we've, you know, we've so we're still in the pilots phase of this, but um, you know, so far we've seen a lot of success with this, you know, a lot of um, you know, these are showing up a lot, you know, getting a lot of citations, you know, these are even getting some click-throughs through, you know, ChatGPT and LLMs.
• 9:09
  SPEAKER_01: And um, you know, this is information that historically I think companies wanted to keep really um private and for good reason, for good reason.
• 9:16
  SPEAKER_01: But now, you know, in an in an era where you know people realize like, well, you know, you can search with more than just keywords, you know, people just go to Google now, you know, people go to even just AIO reviews, you know, and search.
• 9:28
  SPEAKER_01: So like we need to be cited and we need to we need to show up in a way that I think you know wasn't you know wasn't part of the marketing funnel before.
• 9:36
  SPEAKER_00: Right.
• 9:37
  SPEAKER_00: So my question with all of this, like I know a lot of brands have used gated content to build those email lists.
• 9:43
  SPEAKER_00: How is that working now that you've ungated the main ideas, but you still have part of that?
• 9:49
  SPEAKER_00: Are you still collecting emails that direction too?
• 9:51
  SPEAKER_01: Yes, yep.
• 9:52
  SPEAKER_01: So we'll still maintain the gate for the full asset.
• 9:55
  SPEAKER_01: So, like, let's say I'm you know, just kind of spitballing on an example here.
• 9:59
  SPEAKER_01: Let's say like you had a really comprehensive customer story and like you know how how this customer um, you know, work with you on you know this particular product or solution.
• 10:08
  SPEAKER_01: Um, you know, you might want to ungate the main ideas, you know, you might want to throw out some statistics from that so that you could say, like, oh, you know, well, Informatica, you know, caused a you know, 30% increase in X, and like, you know, and here's how we did it.
• 10:21
  SPEAKER_01: And maybe, you know, you could say, well, if you want the full details, if you're really at the decision-making content, you know, then you can download the PDF and we'll preserve the gate and not give them everything.
• 10:31
  SPEAKER_01: But, you know, you can you can create a, you know, somewhere between 1,000 and 2,000 word, depending on the topic, you know, executive summary where you're sort of summarizing the main points so that you know you're look, are you going to maintain the same lead flow of lead capture?
• 10:46
  SPEAKER_01: Of course not.
• 10:46
  SPEAKER_01: I mean, that's not realistic, but that doesn't mean you need to totally destroy the gate and you don't need to just ungate everything for the sake of it, but you need to ungate the right things, you know, in the right way, is how I would say.
• 10:59
  SPEAKER_00: Okay, that makes a lot of sense.
• 11:00
  SPEAKER_00: So my next question is like if I'm listening to this podcast, I like to ask questions that I would think like, oh, okay.
• 11:06
  SPEAKER_00: So are you putting those executive summaries straight on the blog, or is there a different section of your website that you're adding those to?
• 11:14
  SPEAKER_01: So on the Informatica website at least, we differentiate between blogs and articles.
• 11:19
  SPEAKER_01: So blogs are considered something that have like more of a point of view, you know, written by a product marketer or something like that.
• 11:24
  SPEAKER_01: These are on our article section, which is like more of an SEO-focused section.
• 11:29
  SPEAKER_01: But I mean, obviously, from a Google's perspective, there's no difference, right?
• 11:32
  SPEAKER_01: It's just slash blogs or slash articles.
• 11:35
  SPEAKER_01: But for us, like we create that differentiation.
• 11:37
  SPEAKER_01: So um we're putting them on our articles on our website.
• 11:40
  SPEAKER_01: So yeah, they're I guess for you know general software websites, yeah, they're being published as though they're a part of a blog, but from the point of view of the company as opposed to like the point of view of an individual.
• 12:41
  SPEAKER_00: Yeah, that kind of brings me to something I've been thinking a lot a lot this week.
• 12:45
  SPEAKER_00: I've been working on updating my own blog because I had a few, I had a few minutes to do that, and I've been preaching it forever.
• 12:51
  SPEAKER_00: Um I've noticed that when I write blog content, it's it's like for the LLMs.
• 12:57
  SPEAKER_00: Like anybody coming to it, they're gonna learn about AI search visibility.
• 13:00
  SPEAKER_00: But for the most part, that's like an SEO play.
• 13:03
  SPEAKER_00: So are you noticing the same thing with your teams that you're working with?
• 13:07
  SPEAKER_00: Like maybe the blog that you said has the point of view is more personalized compared to the executive summaries that you're putting out.
• 13:16
  SPEAKER_01: Yeah, definitely.
• 13:17
  SPEAKER_01: I think the stuff that has a point of view, um, because we have so many different product marketers in different areas, like say master data management, data quality, et cetera, it'll have their specific point of view.
• 13:28
  SPEAKER_01: Um, you know, obviously that's reflected in what the company thinks and all that, but whereas an article is a little more of just like an SEO play, whereas, you know, but sometimes the blogs will, we have a few that have been picked up like really highly by LLMs because they have such a strong point of view.
• 13:45
  SPEAKER_01: So I think, you know, often I I think these days where, you know, back in the day, like I said, with that kind of HubSpot 2011 inbound marketing model, you know, you would write for the head term and then maybe you would, you know, do sort of like the hub and spoke strategy.
• 13:58
  SPEAKER_01: Whereas now I think things that have a stronger point of view that certainly should use data and statistics and things like that, but to make strong points, well, those points can be, you know, LLMs don't know things, you know, they're just parroting things that they find online.
• 14:11
  SPEAKER_01: So, you know, it can sort of adopt that point of view, and then you could, you know, earn a mention or even a citation from that.
• 14:18
  SPEAKER_00: So there's another thing that I want to ask you about that kind of moves away from the PD PDF to HTML project you're talking about.
• 14:25
  SPEAKER_00: Yesterday you made a post on LinkedIn, or I think it was yesterday, could have been the day before, about moving those key takeaways from the bottom to the top.
• 14:33
  SPEAKER_00: Tell me about that.
• 14:34
  SPEAKER_01: Yeah, so that's one thing actually we're trying as a part of this project.
• 14:37
  SPEAKER_01: But um I, you know, I think typically um that's something that I think has been done for some time.
• 14:44
  SPEAKER_01: But you know, there's a there's a big debate, you know, kind of like we were talking about with um, you know, who knows what's working, what's not.
• 14:50
  SPEAKER_01: There's a big debate around chunking, which I I feel is a bit of a ridiculous term, but you know, it's like an you know, it's sort of an engineering term, data engineering, AI engineering term that sort of got put into a lexicon of like GEOs.
• 15:04
  SPEAKER_01: But you know, in reality, all it means is just um, you know, creating extractable insights.
• 15:10
  SPEAKER_01: It doesn't mean that your entire content, you know, your entire uh article or blog needs to be chunked up into a hundred word bits, right?
• 15:18
  SPEAKER_01: But you know, we found at least with preliminary testing, is that you know, some of you know, we want some of these, you know, some of these data points from our ebooks to be cited, right?
• 15:29
  SPEAKER_01: Um, it's probably not going to be cited from the PDF, you know, it'll be cited from a starticle.
• 15:34
  SPEAKER_01: So I think just putting them ahead, because you know, one thing about um AI crawlers versus Googlebot is they're a lot less generous with their time, right?
• 15:42
  SPEAKER_01: Like Googlebot will spend a lot more time on websites relative to AI crawlers, you know, especially Chat GBT.
• 15:49
  SPEAKER_01: So, you know, if you think of it as like maybe you're only getting a couple of seconds.
• 15:53
  SPEAKER_01: So, you know, if you want to kind of kind of almost like journalism, right?
• 15:56
  SPEAKER_01: Like you, you know, you put your main ideas like right up front.
• 15:59
  SPEAKER_01: Um, but yeah, kind of having almost like a reverse table of contents, if that makes sense.
• 16:04
  SPEAKER_01: It's like, oh, well, like maybe this is stuff that historically might have gone in the conclusion, right?
• 16:09
  SPEAKER_01: And like now we're just putting it at the top, sort of in bullets, you know, just to see like, okay, you know, it's not just a collection of stats, but like, you know, here are the five main points you're supposed to get in the article.
• 16:19
  SPEAKER_01: And frankly, it's good for humans too, because I think, you know, people bounce off websites a lot.
• 16:24
  SPEAKER_01: You know, people are, I think, have been retrained now by Google with AI overviews to not visit websites.
• 16:29
  SPEAKER_01: And, you know, there's a reason that, you know, clicks are down 40 to 60 percent, because I mean, look, as uh an SEO, as an online marketer, you know, AI overviews are a little bit of an existential threat.
• 16:42
  SPEAKER_01: But as a user, you know, it's better, right?
• 16:45
  SPEAKER_01: Like it, you know, instead of you know having to click through a few websites, now you can get your answer in five seconds.
• 16:51
  SPEAKER_01: I think a lot of people don't know that this isn't always accurate and things like that.
• 16:55
  SPEAKER_01: You know, I've certainly had issues um with with different people who have you know gone to these platforms and have you know cited things to me and I'm like, where'd you get this?
• 17:06
  SPEAKER_00: Yeah, I really love it when they just make stuff up entirely.
• 17:10
  SPEAKER_00: Like that's oh yeah.
• 17:11
  SPEAKER_01: Yep.
• 17:12
  SPEAKER_01: There was uh, I think a New York Times report that came out recently that they said like 10% are completely made up and about 50% have significant errors.
• 17:21
  SPEAKER_01: So like that's not you know, if it's something, if you're just looking up something minor, whatever, if it's not 100% accurate.
• 17:28
  SPEAKER_01: But if it's like, you know, important, like uh, you know, kind of like medical advice, financial advice, things like that, yeah, it it might make sense to click through.
• 17:36
  SPEAKER_00: Yeah, that would make sense.
• 17:38
  SPEAKER_00: So speaking of of the New York Times, I was listening to their headlines podcast last week, and uh, they mentioned something in there about politicians who have turned to AEO to correct their information that comes up in the AI overviews.
• 17:51
  SPEAKER_00: Are you tracking how quickly some of the stuff that you're working on shows up in these answers?
• 17:56
  SPEAKER_00: And I'm asking this because the New York Times cited an article, and I'm not sure which one, that the company this politician was working with mentioned the updated information hit the AI overview or whichever chat bot they were looking at within 12 minutes, which is much quicker than I have seen.
• 18:13
  SPEAKER_00: I've clocked it at two hours.
• 18:14
  SPEAKER_00: Are you tracking the timing of any of that?
• 18:17
  SPEAKER_01: Wow, yeah.
• 18:17
  SPEAKER_01: I haven't looked specifically at the time, you know, we'll we'll obviously look at it regularly, but not not within like 10-minute increments or anything like that.
• 18:26
  SPEAKER_01: But I mean, for politicians, that makes a lot of sense.
• 18:28
  SPEAKER_01: You know, if you think about it, like let's say, you know, I I was looking to fill out my primary ballot recently, and I'm like, well, I don't really know anything about a lot of these candidates.
• 18:37
  SPEAKER_01: And it's like, yeah, I mean, I could go to their individual websites and compare and things like that, and you know, what what their political views are relative to mine, but like that takes time, you know, and also like you you might not um you might not know in advance even who's run, you know, depending if you're doing a mail-end ballot or whatever, you know, especially in a a primary.
• 18:57
  SPEAKER_01: So it makes sense that you would just go to Chat GBT and AI overviews and say, like, what are the political views of this candidate?
• 19:03
  SPEAKER_01: And it would just summarize that.
• 19:05
  SPEAKER_01: And maybe it's totally wrong, you know, or maybe it's saying things that are wrong, or like, you know, maybe a candidate's being associated with, you know, something that they haven't endorsed.
• 19:14
  SPEAKER_01: Um, and that's a real problem because that's you know, I mean, we're not talking about, you know, oh, is your software company showing up?
• 19:21
  SPEAKER_01: Now we're talking about democracy.
• 19:22
  SPEAKER_01: Right.
• 19:23
  SPEAKER_01: Right.
• 19:23
  SPEAKER_00: So which is not an angle.
• 19:25
  SPEAKER_00: I mean, I'm talking about AI search optimization all day, every day, but it that's not an angle that I thought about because I'm mostly like software, software, like enterprise.
• 19:36
  SPEAKER_00: That's what I'm thinking about.
• 19:37
  SPEAKER_00: So it was really, it was really interesting to hear that.
• 19:40
  SPEAKER_00: Um, but thinking through all of this with AI search and SEO and then ungating all of your content, well, not all of your content, but most of your content, what is something that has really surprised you about this process?
• 19:54
  SPEAKER_01: I think that um, you know, a lot of doing this at, you know, in in complex uh, you know, B2B organizations is um, you know, stakeholder alignment and you know, making sure people like understand the stakes and like why things are happening.
• 20:10
  SPEAKER_01: And I've definitely noticed people are a lot more amendable to your explanations and your expertise than I think you think they are.
• 20:18
  SPEAKER_01: Um, at least that's been my experience, is like I haven't really, you know, at least when we were Informatica at an independent company, I had no problem getting the ear of like the CMO and or SVPs.
• 20:29
  SPEAKER_01: And even at Informatica or you know, at Informatica under Salesforce, um, you know, there's we we have a strong point of view on how to do things.
• 20:37
  SPEAKER_01: And you know, even Salesforce themselves takes this really seriously in the market in terms of like how you know how they're showing up and AI answers.
• 20:44
  SPEAKER_01: So I think it's I I've honestly been quite heartened to see, you know, how organizations are responding to this, I think, in a pretty agile way, in a way that I think I wasn't expecting, because you know, usually companies move really slow to adapt things.
• 21:00
  SPEAKER_01: And you know, my experience has been the opposite that you know, people really, you know, just like a politician is gonna care how they show up, you know, software comp, you know, big software companies were this can affect your stock price, you know, if you have um, you know, so many a lot of misinformation out there about you.
• 21:17
  SPEAKER_01: At least I know for um Informatica, um, a big problem with us in AI answers was, you know, we were seen as kind of like an outdated solution in the market, right?
• 21:27
  SPEAKER_01: And um, you know, that's because, well, it's a 30-year-old internet company, which is super unusual.
• 21:33
  SPEAKER_01: But, you know, we had gone through a whole um cloud modernization effort, you know, getting off our on-premises solution to um our IDMC intelligent data management cloud solution, but uh, you know, LLMs didn't necessarily recognize that, right?
• 21:48
  SPEAKER_01: And like, you know, is this an answer problem or is this a business problem?
• 21:52
  SPEAKER_01: It's really fundamentally a business problem, right?
• 21:54
  SPEAKER_01: Because that perception in the market already existed and the LLMs are just parroting that perception.
• 22:00
  SPEAKER_01: So, you know, given that I've I've found if you're able to frame things that way, well, people really care about how they're showing up if they're showing up incorrectly.
• 22:10
  SPEAKER_01: Because then, you know, from a sales perspective, well, you know, a sales manager or someone like that is going to come into a meeting and you know, someone might have checked out an AI overview and they said, Hey, like I saw you were not the best solution for this.
• 22:23
  SPEAKER_01: You know, tell me about that.
• 22:24
  SPEAKER_01: And it's like, well, already they're fighting, you know, as a result of this AI overview, or you know, let's say they went to Chat GPT and an analyst ran a comparison for them, you know, for well, you're they're now they're fighting an uphill battle, right?
• 22:36
  SPEAKER_01: In a way that you know wouldn't have happened in you know a non-AI search world because they would have had to look at your website, you know, your competitor's website, maybe they'd look at G2 or another, you know, um comparison, but that that's really all they had access to, right?
• 22:50
  SPEAKER_00: Mm-hmm.
• 22:51
  SPEAKER_00: So so for Informatica specifically, did you have to do any like repositioning of the brand, any of that kind of stuff before?
• 22:59
  SPEAKER_00: Yes.
• 23:00
  SPEAKER_01: We so a lot of actually what we found is huge website, of course, because it's been around for 30 years, you know, probably between all of our you know support pages and knowledge base pages, you know, there's hundreds and hundreds of thousands of pages.
• 23:14
  SPEAKER_01: So, you know, we found wow, there's a lot more references to our old software than there probably should be.
• 23:21
  SPEAKER_01: I mean, for historical basis, you know, you have to have documentation and things like that.
• 23:25
  SPEAKER_01: But, you know, there there was a page, for example, about like how to download Informatica Power Center, which uh Is not a thing you can download, like maybe it was a thing you could download 10 years ago um locally, but it's certainly not something since you know the cloud modernization journey that you can download.
• 23:41
  SPEAKER_01: Not something that was really showing up much in from a search perspective, right?
• 23:47
  SPEAKER_01: Like it wasn't getting a bunch of clicks or anything.
• 23:49
  SPEAKER_01: But you know, it was we found well, LLMs were referencing this page a lot.
• 23:54
  SPEAKER_01: So like what we did instead of deleting the page, because that doesn't really make a difference, we totally revamped the page.
• 24:01
  SPEAKER_01: You know, we talked about like why you actually can't download Informatica Power Center and like, you know, what you can do from a, you know, why you should move to IDMC and you know, all about our cloud modernization journey, you know, adding FAQs, you know, adding, you know, internal links, you know, updating a lot of things.
• 24:18
  SPEAKER_01: And lo and behold, we were able to change that perception in AI overviews, you know, which is connect to the internet really quickly, you know, may maybe not in 10 minutes, but you know, um, I would say pretty quickly.
• 24:30
  SPEAKER_01: And like that, that's just a good example of um that's just one example off the top of my head of things we're able to do to sort of you know set the record straight, so to speak.
• 24:39
  SPEAKER_00: Was that was that the first step that you did before doing the PDF?
• 24:43
  SPEAKER_01: Yeah, the PDF is more of like a project we're working on now, whereas that was something when we were an independent company.
• 24:49
  SPEAKER_01: You know, now under Salesforce, you know, things are um a little different.
• 24:53
  SPEAKER_00: So Okay, there are a few things I want you to take away from this conversation.
• 24:59
  SPEAKER_00: First, before you create more content for AI Search specifically, look at what you already have.
• 25:05
  SPEAKER_00: If some of your best research, your customer stories, your data, or your expertise lives inside of gated PDFs, ask yourself how much of that information is actually accessible to the engines your buyers are using to research you.
• 25:22
  SPEAKER_00: That doesn't necessarily mean go ungate everything.
• 25:25
  SPEAKER_00: Daniel's team is testing a hybrid approach, which means they're taking the main ideas and making them accessible in an HTML article or executive summary, and then they're preserving the full asset behind the gate for people who actually want to go deeper.
• 25:41
  SPEAKER_00: Second, look at what's already on your website, especially the older stuff.
• 25:46
  SPEAKER_00: One outdated page may not get meaningful Google traffic anymore, but that doesn't mean an AI Engine isn't finding and using it.
• 25:55
  SPEAKER_00: Daniel shared a great example of an old Informatica page that was reinforcing an outdated perception of the product until his team went and updated it.
• 26:05
  SPEAKER_00: Third, make your most important ideas easy to extract.
• 26:09
  SPEAKER_00: Daniel talked about moving key takeaways toward the top of content rather than putting them at the bottom in the conclusion.
• 26:16
  SPEAKER_00: That's really useful for AI crawlers, but it's also useful for the person reading who just wants an answer without reading 2,000 words first.
• 26:26
  SPEAKER_00: And finally, pay attention to how AI engines describe your brand, not just whether they mention it.
• 26:32
  SPEAKER_00: If Chat GPT, Gemini, or an AI overview consistently describe your brand as outdated, expensive, difficult to use, designed for a different customer, or anything else that doesn't match your current positioning, don't treat that as an AI problem in isolation.
• 26:51
  SPEAKER_00: Go figure out where that perception is coming from and then work to fix it.
• 26:55
  SPEAKER_00: Because as Daniel said in this episode, sometimes the LLM is isn't creating a perception, it's just reflecting a perception that already exists in the market or one that your content is still reinforcing.
• 27:09
  SPEAKER_00: Okay, that is it for this episode.
• 27:12
  SPEAKER_00: Before I let you go, if you want to go deeper on any of this, the visibility report goes out by email every Friday.
• 27:18
  SPEAKER_00: So where I do the longer analysis, it's less news, more hey, here's what this actually means for your strategy and what you should do about it.
• 27:26
  SPEAKER_00: If you're not subscribed yet, the link is in the show notes.
• 27:29
  SPEAKER_00: I will see you in the next episode.
• 27:31
  SPEAKER_00: Until then, stay visible.