How Conversational AI Is Changing Discovery for Brands and Media

This episode of BIA’s Leading Local Insights Podcast explores how conversational AI is reshaping search, discovery, and online visibility for brands, broadcasters, and media companies. BIA’s Celine Matthiessen and media analyst Mike Boland discuss the shift from traditional keyword search to AI-driven discovery, where consumers increasingly ask detailed questions and receive synthesized answers instead of scrolling through search results.

Following BIA’s recent podcast on AI advertising inside ChatGPT, they examine how conversational AI is transforming SEO, visibility, and content strategy through AI optimization (“AIO”), “zero-click” search behavior, and the growing importance of trusted local content, reviews, citations, and topical authority for broadcasters and media brands.

View the Video Podcast here.

Transcript

Celine Matthiessen 0:00

Hello and welcome to BIA Advisory Services Leading and Local Insights podcast. My name is Celine Matthiessen and I'm the VP of Insights and Analysis here at BIA. Today we're talking about how AIs are shaping search and what that means for anyone who lives and dies by visibility, whether that's SEO media or marketing. The place where opinions are formed and shortlists are created is moving from search results page into AI-generated answers. For brands, the questions become less what's our ranking and more? Are we part of the answer mentioned, cited, recommended, or even sponsored?

Why AI Changes Search Visibility

Celine Matthiessen 0:34

So to unpack what that means for SEO content and media strategy, I've invited my colleague Mike Boland, who's been deep in the future of search and AI driven discovery. I can't talk today. Uh to get us started, Mike. How are you today?

Mike Boland 0:50

Hi, Celine. Good to be with you.

Celine Matthiessen 0:52

Great. So, Mike, when you look at how people search and research today compared to a few years ago, what feels fundamentally different to you?

Mike Boland 1:00

Yeah, I think that that's really where it starts. Uh, you mentioned a lot of things about how all of this impacts like the online visibility playbook. But I think like before getting to that, it's all a function of like how the user behavior is changing. And to answer that question, it seems that we're getting into more of a modality in in search that the industry has been driving towards like for a while and has really just like accelerated

Natural Language Search Takes Over

Mike Boland 1:27

and gone into hyperdrive in the last like one or two years, which is natural language search. We've been talking about this for years, and I think it's the onset of AI engines and LLMs like ChatGPT and others that have one made this possible, and then two, have started to condition the behavior and users that this is now possible. So you see more and more people searching in more evolved ways. And and evolved is the key word because the sort of like the joke always in search, and especially in local search, has been that we we talk like cavemen when we when we search. It's like sushi near me, right? And and now we're using full sentences, we're we're getting into more granularity and we get into that and what that means. But I think that one data point that I think underscores a lot of this is that Sochi, um, a company that works on reputation management and a bunch of other things, a company you know, Celine, um, came out with their local visibility index recently. And as just one of many data points in this report, they reported that um search queries, traditional search queries on Google average four words in length, right? Can you guess the average word count of a typical AI query, whether it be Google AI mode or like ChatGPT? Like guess guess what the the length, the average length in words?

Celine Matthiessen 2:45

In words, maybe double or triple that? I don't, yeah.

Mike Boland 2:49

So it's the according to their study of 23, which is almost six X. That's almost six X, the four words of search. So that that should quantify it. But yeah, that's playing out in a lot of interesting ways.

Celine Matthiessen 3:01

So yeah, so is there anything you can add to that? Uh that's a great observation, high-level observation on um the shift from searching and clicking to asking and getting answers. Any anything else that you can that you can observe or that you've observed that we need to keep in mind?

Mike Boland 3:16

Yeah, yeah. Let's let's like color that with an example, right? Like the caveman search versus the new like natural language search, which makes a lot of sense because it's the way that we think naturally, it's the way humans think. Like it's just it's a return to just you know something that's more intuitive for us. So, anyway, let's take a local search example because that's our world, right? The the sort of you know, coffee near me. That's that's the traditional search. That's three words. Google, I mean, So she's average, they said before, but let's just take that typical search, coffee near me. The way that that might translate today in the way that users are being conditioned and the

Longer Queries And Real Intent

Mike Boland 3:50

engines are being capable of processing, is something that's more along the lines of I'm driving down the 405 freeway in Los Angeles, heading towards Los Angeles. I'm looking for a place to get coffee that is less than one mile off the highway, and I prefer Pete's to Starbucks, and I want one that has a bathroom I could use. Right. It's like that's the new search, not the coffee near me. Um and I think that again, that has a lot of implications for you know the the art of SEO or what's becoming known as AIO, AI engine optimization, that's based on just those new ways that people um are searching.

Celine Matthiessen 4:25

That's crazy. Yeah, um I could see I can see where that that makes a big difference. It's also very helpful because it's very personalized, right?

Mike Boland 4:33

Yeah.

Celine Matthiessen 4:33

Um when so when you look at the tools that consumers are using, like ChatGPT or Google's AI overviews, uh, what do you think they change most about the way people research products or make decisions?

Mike Boland 4:44

So one, it's that natural language. Um, and I think that that again is just more natural um to the way we think. Um, and and that's sort of slowly, or actually I should say it's quickly assimilating. Um, you know, it's gonna take a while to untrain the world on the way they used to search, but that's happening faster and faster. So so it's that natural language, but also I think something interesting that's like another layer

Memory And Dialogue Replace One-Off Searches

Mike Boland 5:07

of advancement on top of that is something we talked about on our last podcast about um ChatGPT's ad program or developing ad program, which is that a lot of these AI engines, including Google AI mode and their overviews, sort of have a memory. So you can have a dialogue that remembers, like the engine remembers that, you know, the conversation you had three weeks ago, or just basic stuff like the fact that you have two kids, a dog, and a nut allergy, right? It'll incorporate those things. And I think on top of that, the thing that's interesting is the capability to make it more like a dialogue. Like if you picture the old search world, you're you're researching a product, right? And you you ask a question, and then you sort of have another question, maybe a follow-up question. You start a new search with that follow-up question. In this way, it's just these these questions and answers build on each other just like a human dialogue. So these multi-part questions. That's that's sort of a hallmark of Google's AI mode, and Chat GBT can do it. So it's sort of this more of a back and forth um that builds on itself as opposed to these isolated one-off queries that you might do 20 of them in the same process of researching like a new TV.

Celine Matthiessen 6:19

So it solves the time consumption in queries, right? But like what other problems are AI solving um that search results never quite nailed? You know, because that's a good example. Is there other ones out there? Like what about like for visual or in the you said it conversations and the deep discovery?

Mike Boland 6:36

Um I I think the uh visual is actually a great example. So visual search is something that again is not new, like a lot of this isn't new, it's just really accelerating and um getting more capable, so it's getting a lot more attention. But visual search, such as Google Lens, has always been something where you hold up your phone and you take a picture of something, and it's great with like if you're on a nature hike, right? It's it's like plants and animals and pets. Like, what kind of dog is that? But I think that it's being incorporated in in ways on the desktop more where it's like shopping-related searches, and like I'm looking for something and you see visual results, and you're

Visual Search And Multimodal Shopping

Mike Boland 7:10

saying you can say something like, I like that, but does it come in blue? Like it's it's it's multimodal in that it's both text or voice input, but also visual. So I think that multimodal capability just opens up new dimension, right? In the way that we search. So we're not always describing things. And for certain types of searches, the ability to search with visuals as opposed to words makes sense, like um, you know, fashion, right? Fashion's a good example, or probably a lot of other examples. Yeah, totally.

Celine Matthiessen 7:38

I mean, for local, I mean thinking auto is one of the top categories and spend for auto, um, um for local, sorry. Yeah, yeah. Okay, cool. Yeah. So do you think this just changes how people? I mean, we talked a lot about how people have changed their behavior um with search. Um, but does it also change how quickly they move from that uh discovery research to decision?

Mike Boland 8:01

Uh totally. I think that that that we talked about this last last week too, or at least alluded to it. The that that convoluted path to purchase that we've always looked at in in local search and local media, where like people are weaving in and out of different devices and screens. And you know, there are some ways that that could be compressed in AI search in in terms of like there's again, it's that conversational just back and forth, and you just arrive upon an answer sooner with less steps and less you know things. But also the fact that there's less of a click stream where

The Funnel Compresses And Clicks Drop

Mike Boland 8:35

the old model, the Google search engine result page SERP model, was all about links and surfacing links, and then you go to that link and oh, that didn't have what I want. We start over, and you're now it's just it's all in that same UX, that same answer. It's all right there. And that's a good thing and a bad thing. We talked about last week how the performance-based, click-based model of online advertising that we've come to know and love is sort of eroding because there's less click-throughs, right? The answer is right there. So I think that there's a lot of implications for marketing. But to answer your question, yeah, I think it could make a more compressed purchase funnel in a lot of cases.

Celine Matthiessen 9:12

Didn't we also talk that there'd be we were talking, I think we were brainstorming last time about new measurability for that too, for that behavior, right?

Mike Boland 9:19

Totally. Like how do you measure it when there's no click? That's an open question.

Celine Matthiessen 9:22

Right, exactly. So when you talk about AI search and AI overviews, how do you define those in in practical terms for an SEO team? So getting into the SEO side of this.

Mike Boland 9:33

Yeah, let's transition into that, like what this all means for the point of marketers and SEO people, totally. So, as you can guess from everything we just talked about, the online visibility playbook has changed. So the the art of SEO has always been about reverse engineering the algorithms, Google's page rank algorithm. What are their ranking factors? How is that changing? It was always a moving target. Now, like that's still there, that hasn't gone away.

AIO Vs SEO And What Overlaps

Mike Boland 9:58

But now there's also this like this sort of new entity out there that's providing answers, and that's all the stuff we're talking about. And similarly, it's going to be about reverse engineering the ways that it finds information, the way that it surfaces answers. And unfortunately for marketers, there's only some degree of overlap between the two. Like, there's been this sort of uh uh ongoing question of like, if I was on top of my SEO game, does that automatically put me on like a head start or higher footing in getting surfaced in AI engines? And the answer is like sort of. Um, I'll mention that SOShi study again, their local visibility index. They they basically measured a bunch of multi-location brands, right? And their search engine visibility and who was like performing best, and then their AI engine visibility, excuse me, and who was performing best. There was a 45% overlap in brands in the top 10 of each of those two respective lists. So that sort of quantifies that like those who are doing well at SEO, some of them are also doing the right things with AIO, as we're talking calling it AI optimization. So you sort of have to do the same things you were already doing with with search and SEO, but also this new set of rules. We can get into what those are, but that's just like an overview on like what these marketers are now faced with in having to sort of like serve two masters and appease the gods of Google and now also the LLM gods.

Celine Matthiessen 11:30

Well, yeah, I mean, because like Google, you know, there's everybody, there's people have built careers off of good SEO, right? And it's not it's the keywords, it's the it's the content, you know, and all those things have rankings, it's reviews, it's new content, original, and AI, AI is that, but more, right? I mean, because they're it's you know, so what what you know what sources are they pulling from that's different?

Mike Boland 11:56

Yeah.

Celine Matthiessen 11:56

Do you know? I mean totally, yeah, yeah.

Mike Boland 11:58

Let's let's get into a few of the different I don't.

Celine Matthiessen 12:00

I I I kind of know. I think we talked about like I know that they're pulling from um, you know, more from video because it's easier. Um, that that's an easy one. Um, so because they because Google does the transcripts and they could they look at that, um, they don't actually watch the videos for those out there, but no, no, those those transcripts are gold mined for for L.

Mike Boland 12:20

Goldmine for training sources.

Celine Matthiessen 12:22

And there's also, I think there was just a lawsuit too. There's a bunch of educators that are suing Google and YouTube because they scraped for educational purposes their YouTube videos and they're going after them to train their model. And I think we're gonna see a lot more of that, but that's how I know that. So I'm sorry, I digress.

Mike Boland 12:38

But no, no, let's let's go over, let's go over some of those authoritative sources of AI engines. But before we do that, just a really quick, quick note. You mentioned that there are a lot of folks that have like built a career around being SEO experts and consulting brands on how to you know surface higher in Google. I think those folks are actually or the smart ones are actually gonna be elevated in this new world as opposed to like disrupted in a negative sense, because the all the brands out there now have so many more questions than they had before. And I think those SEO folks, the smart ones, are really jumping onto these new ranking factors and how they optimize and all that stuff like that. So, uh, and they're they're gonna be more needed than ever because there's more question marks and confusion than ever. But anyway, this is all a moving target, and it's all about reverse engineering and extrapolation, just like SEO was. So I'll go over a few of them that we sort of know so far, and then I just I preface it that way because this is a quickly moving target. But anyway, let's go back to the example we gave in like the longer, more nuanced local search query. The driving down the highway, looking for coffee, it's somewhere off, you know, within a mile off the highway. Just the granularity in those queries means that these AI engines have to like dig deeper, right? To answer that level of granularity than traditional search ever has. So, where do they find that long tail of granularity? A few places. One is reviews. Reviews are more important now than ever. Yelp reviews, because you

New Ranking Signals Reviews Reddit FAQs

Mike Boland 14:08

have just like these endless scrolling pages of like people talking about these things. So that's where the LLMs are going to train themselves to get these answers and synthesize the answers to these questions. So reviews is more important than ever. So the implication for marketers is it's always been pretty important to work on your reputation management and manage your reviews and respond to them and cultivate them and all that good stuff. That's now more important than ever. Another one for a very similar reason of that long tail content with all that granularity is Reddit. Reddit is the new golden goose of LLM, you know, this AIO, this developing art of AIO. Be on Reddit. Make sure you're there, make sure you have an account and cultivate your brand and mention your brand and monitor it and all that good stuff. Um, another one, again, based on the ways that people are searching because they're they're asking questions instead of just typing in keywords, is FAQs on your website. This is a low-hanging fruit one. It's harder to like get people to write good reviews and get people to talk about you on Reddit. Low-hanging fruit, just put an FAQ on your website. And that sort of helps um the engines like see, oh, there's a question and an answer. We can we can work with that because that's the same way people are searching. Um, and then another one I'll mention is citations or back. Citations used to be a very um you know prominent Google ranking factor. And then there were a number of years where all the SEOs and Google was saying, like, you know, that's not as important anymore. And what I mean by citations is like, make sure you're mentioned by the trade press or make sure you're on like the local chamber of commerce membership website. Like Google used to treat that as authority, and now they they sort of pull back from now. That's now more important than ever. Um, and and one like tactical takeaway I heard recently is like go back to issuing press releases, issue press releases, you know, on like free wire services like PR log or other or just on your website. Um, so that's another one. Um, press releases, and and then I'll I'll mention something that is actually that LMs are not paying attention to that that Google has more importantly. So go uh Celine, you know this in in the past I don't know, five years or so, Google's been paying much more attention to technical SEO. So there's on-page SEO, which is the words you're saying, the keywords, the FAQs, all that. And then technical SEO is the stuff in the background that's like, do you have the proper schema markup? And are you utilizing your canonical tags strategically and page meta descriptions and core web vitals? That's another one that's really accelerated in the last five years. How fast is your page load? Properly sized images that load fast. Um, LMs don't care about any of that. So technical SEO is out. And I shouldn't say out because Google's still there, right operating as Google does. But for this half of the searches that are happening on not half, I don't know what the breakdown is, but for the portion happening on AI engines, technical SEO is out, but all those other things that I mentioned are reviews, Reddit, FAQ, citations, and and that list just keeps growing.

Celine Matthiessen 17:09

That's crazy. Well, I mean, it kind of makes sense because the technical is really there to kind of check off the boxes to say that this is an actual, you know, live working, breathing, current, you know, business farming product, not content farm. That was to fight the content farms. And yeah, so that's very interesting. So speaking of like FAQs, uh, are there particular query types? We talked about one, you know, getting coffee and being driving query types or verticals, um, where you're seeing that AI style answers take over faster, like or you know, because of consumer behavior or just because of the nature of what um of AI and what um what it's being used for, populari, you know, popularity.

Mike Boland 17:52

Yeah, I think popularity is the the right word because it sort of mirrors everything that's out there. Like, you know, LLMs train themselves on what's they're only as good as what's out there. So if you look at the different types of content out

Which Verticals Win First And Why

Mike Boland 18:04

there that are more prevalent than others, I think that sort of answers your question in terms of like what they're doing a better job answering, what people are asking more. So you can think like, you know, sports, news, stocks, politics, um, they're all going to be rich sources of information. And I think that's a self-fulfilling prophecy because the more people ask those questions, the more answers are surfaced, the better the engines get because they learn, obviously. So I think that that that's the case. And then on another level, I think that there's, you know, that's just more of like a passive reality, like that's just inherent. But I think there's also some ways that these engines are sort of trying to manufacture some competency and specialization in different verticals. And we talked about them on the last podcast, which is um healthcare. Um, OpenII sees healthcare as like an area that is so rich in terms of specialized knowledge, uh, but also a very high value vertical. Another one is legal. Um, I read the other day that um Anthropic is building a sort of variant of Claude that's all about the legal vertical. And that one makes a lot of sense because like you picture law firms, you know, half their staff has like all these paralegals running around and looking up stuff and all that, and then a lot of that can be automated. They see a huge opportunity there. So I think it's a combination of just like the inherent share of topics on on the broader web, like that's gonna sort of mirror what what's happening in these LLMs. But um on another level, there's some deliberate maneuvering by some of these AI giants to sort of cultivate some of these high value verticals.

Celine Matthiessen 19:39

Right. So retail, of course. I mean, we're seeing that all over the place. Totally. That's a big one.

Mike Boland 19:43

Shopping, I think, yeah, absolutely.

Celine Matthiessen 19:45

Shopping is huge. Yes. Yeah.

Mike Boland 19:46

And also these ones we're talking about, I think I mentioned this for for healthcare, but these are all high vol, these are all monetizable verticals. So they're they're they're like salivating, not just in the they're big, but they're also high value because the commercial intent on a on a shopping search is gonna be greater than like the type of research you're doing for like a research paper or you know, stuff like that. You're not necessarily gonna buy something, you're just looking for knowledge.

Celine Matthiessen 20:10

Right. Yeah. So from an advertiser's point of view, uh what's just what is the strategic appeal of answer adjacent placements, you know, and should you know, do they offer anything like it anything? What's the value for them to kind of really invest in this?

Mike Boland 20:29

Yeah. So um there are some signs, some evidence that the traffic is actually converting better, like traffic that comes from an AI agent um as opposed to traditional search. Um, and it's sort of a double-edged sword. Like you've probably heard all of the publishers

Ads Near Answers And Better Conversions

Mike Boland 20:46

and brands and anyone with a website sort of lamenting over the last year or more about the fact that their traffic has gone down because of the AI answer. You know, like we talked about earlier, there's less clicks happening because the answer, the question's answered right there. And so um, they're getting less clicks, but there is some evidence that the clicks they're getting are higher intent and better in terms of converting better. So I think this study came from Adobe, Adobe Shopping. Um they they found that their AI traffic or AI traffic in the sites that they were monitoring converted 42% better um than non non-AI traffic. And interestingly, a year ago they did that same study and it was 38% worse. Oh so it's it's like a massive difference in, I guess, just like the the way that people are starting to utilize these engines for shopping and for things that are converting. So I guess that's good news to answer your question about like how should brands think about prioritize like putting themselves there, um and and like paying attention to the everything we just mentioned, the AIO playbook. Like, not only is it important, but to underscore that it's high stakes and it's a missed opportunity, maybe if you don't go after it.

Celine Matthiessen 21:59

Right. I mean. your c competition could be right. So it's it's like yeah yeah. So you know we talked about um a couple changes. Is there anything else that a brand um or marketer should prioritize in the next six to twelve six to twelve months for SEO strategy against AI answers. Anything can you predict the future? Anything well I would say or make a gas a good educated based on all your knowledge and experience?

Mike Boland 22:26

Yeah I would answer it this way. The the stakes I just I think I just said high stakes, right? The the the degree to which all the stuff we're talking about, how how heavy it is, how much gravity it actually has is only going to grow because it is a direct function of the number

The Next 6 To 12 Months

Mike Boland 22:43

of searches that people are doing on like they're turning to chat GPT instead of a Google traditional search. The more that happens the more market share that AI engines gain in terms of like query volume, the more important it's going to be to pay attention to all this stuff. And I think that's only going to grow. So I think like in terms of the how and something more concrete or tactical, I would say that marketing teams need to learn this stuff, become experts in AIO, play with it, learn by doing I think that um you know that should be just an extension of the things you're already doing and and don't ignore it. You know, don't let it uh catch up with you. I think that's very important because it's happening so fast.

Celine Matthiessen 23:26

Right. Yeah. Well and talk to your SEO consultants too because they're on top of it apparently right they are yeah so well Mike thank you so much for being here today and sharing your thoughts and insights we always appreciate I you know I do we always appreciate your value and time with us right um so anything any other parting knowledge are we good um no no this was fun thank you thank you for having me and what I'll say is like to echo what I just said which is this is moving so fast like we could do another episode in two weeks on the same exact topic and probably have like a new list of you know ranking factors.

Mike Boland 24:00

So so this is going to be moving quickly and we're gonna stay on top of it.

Celine Matthiessen 24:03

Yeah I think I you know I would love to have you back. I think we should you know circle back on this in a couple um in a couple weeks or a month or so and see what's changed. That would be a great follow-up to this and for everyone out there listening thank you for joining us and we might have another one coming soon. Um we hope you're enjoying our podcast this year. And if you have any topics or you want to see us talk more about this to suggest or you want to come on in as a guest, reach out to us at podcast@bia.com. We look forward to bringing you more insightful conversations on local media throughout the year ahead. Thanks