Glad to have you with us on The Search Session. I’m your host, Gianluca Fiorelli, and for this session, I'm joined by Abby Gleason, an SEO product leader based in Greater Minneapolis, USA.
There’s always something new to chase in search, and AI has only made that list longer. Abby and I talk about staying curious, choosing where to put your energy, and building a brand people trust. Along the way, we get into useful experiments, stakeholder conversations, and a few AI search myths.
What you’ll learn in this session:
How to measure AI search visibility and impact: combine tracking with customer surveys, ground prompts in real customer questions, and assess visibility across broader topics.
Why SEO strategy needs to extend beyond your website: customers search across multiple channels, making relevant content and performance tracking across the ecosystem essential for your brand’s visibility.
Why traffic alone no longer measures SEO success: conversions, revenue, branded searches, and direct visits help reveal impact as search journeys change.
How smaller brands can build influence in search: focus on a clear niche, earn mentions in sources that inform AI answers, and create a product and customer experience worth talking about.
AI’s role in enhancing SEO skills: AI can support research, data analysis, and brainstorming, but SEOs need firsthand knowledge and experience to challenge its recommendations and make informed decisions.
What makes AI worth pursuing: avoid AI shiny object syndrome by solving real user problems, testing ideas, and using evidence to guide stakeholder decisions.
Why abandoning informational content is a bad idea: it limits your influence early in the customer journey and reduces opportunities to demonstrate expertise, earn AI citations, and attract potential customers.
How to win stakeholder support for SEO: run small-scale tests, use data to demonstrate business impact, and connect SEO issues to revenue to secure budget and resources.
Why Reddit mentions aren’t a shortcut to AI search visibility: earn genuine recommendations through a good product or participate transparently and helpfully, rather than using fake accounts and spam.
Abby brings plenty of hands-on experience and a refreshingly practical take on making SEO work, so join us for the conversation.
Topics covered: search influence · AI search · search beyond Google · brand authority · informational content · SEO measurement · conversions and revenue · customer research · AI tools · shiny object syndrome · SEO experiments · stakeholder support
About the Guest
Abby Gleason
Lead Product Manager, SEO at Compass
Abby Gleason has more than a decade of experience driving organic growth. She specializes in SEO, AI search, conversion optimization, and website experience, working with engineering and design teams to help people find what they need and turn visits into signups.
She recently joined Compass, a real estate brokerage that combines agents’ expertise with technology to help people buy, sell, and rent homes. Previously, she led organic growth at Upwork, which she joined in February 2025, and worked at Scribd from July 2022 to January 2025, managing SEO and acquisition across Scribd, Everand, and SlideShare.
Abby also shares practical insights through articles, case studies, podcast interviews, and speaking at industry conferences.
Video Chapters
Transcript
Full conversation between Gianluca Fiorelli and Abby Gleason.
Gianluca Fiorelli: Hi, I'm Gianluca Fiorelli. Welcome back to The Search Session. Today we are going to have a guest from the central states of the United States, from Minneapolis. I must admit it: for a long time, I didn't really know her, but slowly I started seeing things she was sharing on LinkedIn, things that she was writing. For instance, she wrote a lot on Wix, but she's also writing for Search Engine Land, and I also think for Moz, and so on. And so I started to be intrigued by this person.
She was the senior lead for organic growth at Upwork. She previously worked at Scribd, and she very recently started to work as an SEO lead for Compass Real Estate. This guest, our guest today, is Abby Gleason. Hey, Abby. How are you doing?
Abby Gleason: Hi, I'm good. What an introduction. Thank you.
Gianluca Fiorelli: You're welcome. Very well deserved, no?
Abby Gleason: If you say so. I'm from the Midwest, so compliments we tend to deflect, but I appreciate the kind words.
Gianluca Fiorelli: I'm Italian. I'm Mediterranean and very expansive in my compliments. So, how are you doing?
Abby Gleason: I am great today. I reflect your sentiment about the September weather being gorgeous here, so I can't complain. How are you?
Gianluca Fiorelli: I'm fine. I'm fine. It's still hot here in Valencia, Spain. Well, it's finally not as hot as it was during the summer, but fall is a wonderful season here, and I'm sure that's also true there in Minneapolis. And what about SEO? How is it treating you lately?
Abby Gleason: So despite all of the upheaval that we've seen, I find this, just in the past decade of working in SEO, this is by far the most exciting time I've been a part of and the most energized I feel. I think it's been really exciting to see how search has evolved from 10 blue links and tracking rankings obsessively on Google to "Are you visible across the search ecosystem, from social media to AI search like ChatGPT and now Google's AI Mode?"
It's just a really exciting time, and I think it's also just been very nice not to have to stare at keyword rankings day in and day out and get those emails from clients. So, I'm excited to talk about that evolution. There's still a lot of foundational elements that apply, but there's also new things to consider that have been really fun to delve into, and I honestly see it as a pretty big opportunity for SEO folks.
I see a ton of new job opportunities popping up with AEO and GEO in the title. Candidly, there are lots of AI visibility companies that are looking for SEOs to partner with, so there are lots of opportunities to be a brand advocate and test out new tools. So, long story short, I'm having fun with it.
How to Measure Success Across the Search Ecosystem
Gianluca Fiorelli: Yes, indeed. It's an exciting time. And it's already been four years since ChatGPT was rolled out by OpenAI. It seriously seems like ages ago already, but it's only four years ago. And four years ago, we were talking about Google updates, spam updates, core updates, and so on—and we still are—because classic search is still the main driver of traffic.
So yes, it's surely an interesting time, an exciting time. And you were saying SEO, for certain kinds of things, was becoming a sort of routine, and this is making us feel young again. But you were saying something: we were obsessed somehow with reporting classic visibility. Somehow that was relatively easy. It was complicated by Google with all the SERP features, so ranking number three wasn't anymore like ranking number three ten years ago.
But now the problem is—from some very structural way of measuring the impact of our work—that we may still be in the phase that we don't really know how to measure and attribute to our work the success of a brand in the search ecosystem. And I say a search ecosystem, meaning not only the classic search engines, Google and Bing, but also the LLMs and eventually other places where there is a search bar.
How do you do it? Because you have worked at Scribd, you have worked at Upwork, and you started to work at another big company, an enterprise company. So, during this year, how have you developed your own system for measuring the success of AI?
Abby Gleason: I'll just start by saying, if it feels complex to folks, it's because it is. We've moved into this search everywhere optimization ecosystem in which zero-click or impression-first channels make up a lot of the influence.
So, for instance, I always use this example. I went to Mexico City recently. Five or six years ago, I would've started my search on Google, clicked on something, and then I would've been attributable. You could have tracked my path to signing up for some experience. However, now it was an entirely different experience. I went first to YouTube and TikTok to find what people talk about. Then I went on ChatGPT to create myself an itinerary, and only then did I use Google to book the experience. They have no idea that I was influenced by these various channels.
And so I think the reporting structure and expectations have to be shifted entirely. A couple of changes that I myself have made: I know that these tools are controversial, but we have definitely added in a tool like Profound. We use that to track our top prompts. We gather those from top-converting keywords, so basically the topics that we know drive us revenue, and then we've transformed those into prompts. It's not a perfect science, but it's the best we have. And then we're able to showcase to leadership, “This is our share of voice. This is our visibility. Here are the citations where we're showing up that influence these answers.”
Another element, too, is that I think it's pretty important to try to get qualitative data into your reporting. That's something that I don't see SEOs talking about at all. Why don't we have “How did you hear about us?” surveys either in our registration flows or as a follow-up afterwards? I think if you just rely, for instance, on LLM referral traffic, you're going to think, “Oh, this is a fraction. This is 1% of my total traffic.” But it could be, let's say, 20% of your attributable reasons that people came to your site.
So I think it's helpful to add in a visibility tracking tool in addition to your traditional rank tracking and organic traffic tracking, LLM referral traffic, and then having that human element of, “Okay, where did you actually hear about us?” Because I give it a 99% chance that you'd be undercounting if you rely on quantitative methods alone.
Gianluca Fiorelli: Yes, indeed. Also, because I remember the time when we were also discussing dark organic, dark social, and so on. I paint mini soldiers, and it seems like we are in a grim, dark era of analytics, so not really understandable. But yes, you were really, really right.
Sometimes I see people still talking about a classic idea, which was: SEO should talk with customer care, for instance, or do public surveys with people when entering the website or post-selling in order to understand what kind of questions are coming out.
And actually, what you were saying is that these kinds of real-life polls should be the basis of the prompt tracking that you are going to put, in your case, in Profound. We are in Advanced Web Ranking. Also, Advanced Web Ranking offers the opportunity to track prompts and so many other tools.
And maybe that is better, surely better, than leaving the tool the opportunity to propose its own prompts, basically. And surely better than asking an AI to transform a simple query into a series of prompts.
Abby Gleason: I love that.
Gianluca Fiorelli: But it surely is more interesting. I have to ask you one thing because there are two sorts of schools. The one using, obviously, the prompts, which is fine. Sometimes I'm still skeptical about the quality of the prompts. I don't think they are perfect, but there are certain kinds of methodology that you can use in order to find at least a statistical value for the prompts. And then there is also the way, for instance, Waikay is saying, “Okay, the most important thing is how visible you are for a topic.” So maybe this is also a good idea to balance the two kinds of issues.
Tracking specific user prompts is essential for high-intent keywords, but monitoring topic association tells you how LLMs perceive your brand authority overall. You don't have to choose one over the other: Advanced Web Ranking tracks both sides of the coin.
AWR lets you monitor your top-converting prompts while also auditing the core topics and entities AI models associate with your brand and competitors. Try it free.
Why SEO Strategy Must Extend Beyond Your Website
Gianluca Fiorelli: You were saying Mexico. I want to dig into your Mexican vacation. I guess, I hope, it was a vacation.
Abby Gleason: Yes, it was a vacation.
Gianluca Fiorelli: A Mexican vacation experience, and you said, “Okay, I started with TikTok.” This is a classic already. Remember Google a few years ago saying that the very young are now searching for local search on TikTok and Instagram. It was already there. And then you say YouTube, for video, et cetera, et cetera. How much is this also representative of the fact that an SEO strategy should not be reserved only for the website or focused only on it?
Abby Gleason: Yes.
Gianluca Fiorelli: What is your vision when you are, for instance, working with a client, now with a company, when you are focusing on the big-picture view of the strategy? When do you think it's better to focus more on the outside than on the inside, and how do you make the two work together effectively?
Abby Gleason: This is a great question. I always go back to first, know where your customers are, and this will vary based on industry. But in our case, we had absolute validation that our customers were using YouTube and TikTok to search for the topics that we cared about. And the way we validated both was in traditional ways.
You can easily see YouTube search volume, and we found tons of volume for the topics that we cared about on YouTube. And then you can also just see how many videos, for instance, on TikTok are created about a particular topic. So for Upwork, it's the world of hiring and freelance work, and you can see without a doubt those are popular topics on TikTok.
We should have a voice there, whether it's Upwork as a brand talking about it, us partnering with influencers, et cetera. And so we really validated that approach first. It's not for every brand, but I do think that it's increasing as more and more search and more and more users navigate to these user-generated content sites because I think people see, “This is where I can get more trustworthy, more human reviews and recommendations.”
That's why I think we see Reddit appended to so many Google searches, because people have gotten wise to SEO's tricks of affiliate links and stuffing things in articles. And the days of that being the primary force of influence, I think, are over.
And so that's something we do at Upwork. The website is one part of our strategy, but it's not the whole of our content strategy. We also want to own the whole ecosystem. And so for us, we've identified a few channels: YouTube, TikTok, Reddit, and LinkedIn. We actually have a content strategy dedicated to each of those channels, and then we track our performance in all of those. So, video views—are we ranking for certain terms with our videos?
And actually, YouTube and LinkedIn have been our fastest-performing channels, faster than really any other channel. When you create a video or a post, Google is pulling this in ridiculously fast, sometimes within a day or two, right into the search results. So we've seen a lot of success so far with this method and are only going to continue scaling it up.
Gianluca Fiorelli: Yes, indeed. And everything you are saying is also making me remember how Google, very recently, I think a couple of months now, is giving websites the opportunity to create Google Search Console properties for their own social media profiles, in this case.
Abby Gleason: It's a big signal.
Gianluca Fiorelli: Because it's clear with video, because video, substantially in the classic form, usually is just YouTube. There is more variety in the Shorts box, where you can still see a lot of Instagram and a lot of Facebook, not so much TikTok.
We usually blame Google for being very obscure with everything in Search Console, but at least they did one thing right, giving us the opportunity to see this kind of performance that is really important for us. Because visibility, even if we want to just talk about classic search, has not been just the classic search snippet for a long time.
Abby Gleason: Yes, absolutely.
Zero-Click Search and the Shift From Traffic to Conversions
Gianluca Fiorelli: There are these other features where you need to be visible, and not only when people are doing the first search, but also when they are doing the second search with further searches inside Google, so the classic concept of messy middle.
But now it's expanded to all these many surfaces that we are dealing with. And one thing that is common to all the surfaces is becoming common with zero-click on Google Search. It's evident in LLMs because, sincerely, how many times do we click on a link on ChatGPT? Also, because they are substantially hidden. And social media was never really a place where anybody was going to click. Sometimes the social platform makes it very complicated.
And even if everybody is saying that searches are exploding, which is probably true, it doesn't correspond to a traffic explosion. You have talked and written about this. I saw you writing about or talking about this kind of problem. The importance of the click, the remaining click, and the traffic: how important is it going to be for SEO to pay attention not only to how people are coming to our website, but also to how long and how profitably the people visiting our website are remaining on our website? So the needed collaboration with CRO and usability. What is your opinion about this?
Abby Gleason: In general, I would say I haven't really reported on traffic in any substantial way in years now. I think traffic is sort of the wrong metric to look at due to exactly what you said. Traffic is contracting at what would feel like an alarming rate if people just weren't searching anymore.
But I think, as you said, that's just not the case. If anything, search has evolved and is only increasing. I think the metric to track is conversions. So, in our case, we've seen traffic decline, but we might not have seen registrations, conversions, and revenue decline at the same pace.
And we've actually had to shift what channels we look at to validate that. So instead of it coming through non-brand organic, it might be coming through in paid as people search for just our brand name rather than those top-of-funnel, mid-funnel queries.
We've also started looking more at direct, people just going directly to our site, as well as visits to our homepage and looking at the number of branded queries. So, are people typing in our brand name more and searching for us?
And so we've moved further down the funnel in our reporting structure because traffic is no longer the right proxy to determine if your SEO strategy is working, in my opinion.
Gianluca Fiorelli: Yes, it's a dependable opinion. Surely there are people listening and watching us who say, “Okay, but we also need traffic in order to obtain these kinds of results in conversion.”
How Smaller Brands Can Build Influence in Search
Gianluca Fiorelli: You were saying you started to also track direct traffic, typing traffic, and branded traffic substantially. So in this case, SEO, GEO, AIO, or whatever you want to call it, depending on the surface, is becoming a battle for influence in this sense. How visible is your brand in order to influence and be memorable in the search space? And eventually, the better measure is how memorable it is when attached to the keywords or entities that you want to be associated with.
You have worked with Upwork. You're going to work with an important real estate company. And a classic counterpoint by others could be, “Okay, but these are big brands.” How would you justify that creating a strategy and developing tactics for obtaining this kind of influence in search is also valid for every brand, independently of how big they are?
Abby Gleason: I love this question. Taking Upwork as an example, I feel very fortunate in some ways because Upwork is the best-known brand so far for a freelance marketplace, and so our share of voice and visibility score has always been high. On the flip side, I do work with some clients that have the opposite situation, where we track their key prompts, very obvious, “What is the best XYZ platform?” The number one thing they need to show up for, and it's abysmal.
So I've actually been working with a client recently who falls into the bucket that you shared of a lesser-known brand. My two pieces of advice would be this. One is that I believe in every niche there is a possibility, regardless of how small your brand is, to be the number one in that niche. You may not be the number one for every single super broad term, but if you're going to be a successful business, you need to be the best at something. That might be as small as being the best auto shop for this one specific type of car in your area. But then you have to be the expert on that one thing. It might not be every single type of car, but it's that one car.
So, learning where you can really build authority as a brand. And that's where I caution SEOs when you consider going in-house somewhere or working with a client: really niching down with them and saying, “Who do you serve?” And, “What do you want to show up for, and what can you realistically show up for?” Because I've also worked with clients who want to be number one for everything, and you have to set expectations. It's just not possible. So one is getting really clear positioning around your specific niche and who you serve.
And then the other, where I find tools like Profound are helpful, is you can actually see, for your prompts that you want to show up for, these citation sources that feed into that. And you can see if your brand is cited. This client I'm working with, for their top prompt, they are not cited in eight of the ten citation sources that continually show up.
I'm not affiliated with this tool at all, but AirOps Offsite is something that my team has used in the past. It's really helpful because they can identify, for those prompts, the citations you're not showing up in, and then do outreach on your behalf. You don't need a team to do this. You can do this in-house. You can hire a freelancer or another agency to do this. But I'm really intrigued by these service-plus-software collaborations where they have the software to find these gaps and then reach out on your behalf. Upwork was doing that and has seen pretty quick successes and a pretty good inclusion rate in those citations.
So, long answer: get really clear about your positioning and make sure that you're cited in the sources that reference you.
And then there's a third, less tangible element, which is just building a product that's worth talking about. You can't game things like Reddit and social media beyond a certain point. People are really smart, so treat your audience as smart people and know that if you have clear issues with your product, people are going to sniff that out. You can't control every review and everything people say.
So I think just focus on the key product experience. Customer support: when I see companies stop investing in that, I think that's a mistake. Having good customer support is essential. So that's the third element: have a really good product, and I think you're going to have a better base than the majority of companies.
Gianluca Fiorelli: So, if any of you is responsible for customer care in any company, remember: don't give it to AI. Maintain at least a couple of people personally replying to the emails and answering the sometimes painful calls from customers. But maintain them because it's surely something people want.
Abby Gleason: Can I give an example of that, actually? When we were chatting about what to talk about on this podcast, you mentioned Framer. I recently posted about them. Again, seriously, I'm not paid by them, nothing. I just love them. I was using Framer to build my personal website, which I just relaunched last month, and I didn't realize that I was using their most expensive AI model. I ran out of credits so fast.
And I reached out to their support. I got an actual human, and they gave me more credits right away. Just no question, and I was able to complete my site using a lesser model. It was just one of those things where, if that hadn't happened and I had been hit by an AI bot and they had refused to help me out, I may not have paid for Framer or continued to pay for it. I may not have made that LinkedIn post.
And who knows? That's not necessarily influential, but it's just one of those things where you can see the cycle. That one person that you're paying to help out has created this chain where now we're talking about it on a podcast, and I demoed Upwork on them. You see how it's just this chain reaction when you actually treat people like humans.
Gianluca Fiorelli: And this is also substantially what makes a brand a memorable brand.
Abby Gleason: 100%
Using AI to Enhance SEO Skills, Not Replace Them
Gianluca Fiorelli: Let's talk about how you are using AI as a means, as a tool. You just said that you redesigned your website with substantially AI-fueled and AI-based software and tools. But also, you started creating tools internally when you were at Upwork. Because this is always a big fear for SEOs, marketers, and people in general in digital marketing who are seeing how AI can be somehow a competitor to their own job. So how do you use AI in a way to enhance your skills? How did you test it? What did you find that wasn't really working well, and what did you find that was perfect for having AI as a companion?
Abby Gleason: I also really like this question. For me, I think of the term “human-in-the-loop.” AI can give you a starting point or a middle point, but a human always needs to be involved in the endpoint.
I use it every day, but the way I've used it has really evolved. I use it a lot more to do things like data analysis and finding gaps versus doing things like writing briefs or writing content. I think that's one way SEOs might think of using AI, but we don't really use it for that as much.
And so, if it works, I can just give a couple of examples. Something that I've used AI recently on is as a research agent. I think most SEOs can relate to the fact that it's hard to keep up with the new topics that are coming up in your industry or your area of sales and what to create new pages on. There are various ways to do this. You can use Ahrefs to find what your competitor keywords are. You can talk to your customers, but it's largely been manual.
That was one thing Upwork was really dealing with: there are so many new skills that people can hire for, and we need to understand what those are and create at a more rapid pace. We can't rely on any sort of manual process for this. And so I used an AI agent as my research assistant to basically say, “Look at the pages we have, scan these every week, and then look at the things that people are searching for on our site.” So, are they posting jobs for skills that we don't yet have? Are they doing internal site searches and coming up short? And then, candidly, I also include our competitors’ sitemaps in there. What are they publishing that is a gap for our own thing? And then I have it also crawl a lot of different websites and different searches to see what's happening in the industry.
Every week it's sharing with me, “These are the 20 topics that you should be writing about.” And then we create those pages, and we've seen rapid success with this method. That's where I think AI makes sense: taking something that would've been manual and highly intensive, going through a lot of different keywords and topics, and synthesizing it into something, comparing it with the pages we already have, and then producing an output that a human can take. And so we don't just auto-generate these pages. We look through every single suggestion, and that's where we get a little more manual.
Gianluca Fiorelli: I think this is a very important thing that you are saying. Sometimes the classic problem that I see, for instance, in these wonderful AI-based SEO frameworks, to audit everything from technical to link building to content strategies and all, is where you can see the classic AI slop version of SEO audits. But you said that correctly. In fact, I think that one of the best things to do—AI is wonderful for statistical analysis. And you were saying your AI agent was going to say, “Okay, these are the 20 things.”
Please, everybody doing something with AI, always remember to remind AI not to give opinions, just hints about something to look at better. So you are going to give the opinion. And eventually, you can start brainstorming with AI when you want to find new ideas. For instance, “These competitors are writing about these 20 topics. How can we talk about these 20 topics without sounding the same, without creating the same kind of content as our competitors?” So, the consensus, information gain, and so on.
There can be a place where you can use the AI agent and a different AI agent as your sparring partner or as the ones acting as sparring partners with your creatives.
Abby Gleason: Yes, exactly. I think the key is that AI shouldn't be treated as the boss. And I think that's what happens: AI will give its opinion, and that's treated as fact. “Okay, this is what I have to do now.” In reality, you as the SEO are the boss. You are using AI as a tool to spitball or brainstorm, but ultimately you are the person with taste and experience and have to read between the lines and make that decision.
Because even with this agent, it's suggested things that do not make sense. For instance, maybe we're seeing something increase, but we already do have a page; it's just worded slightly differently. There are all these things that you still have to know.
And that's why whenever I take on any sort of new freelancing client or start a new job, I audit the site manually. It takes time, and it's worth doing so that you have context when AI does spit out recommendations. You can actually challenge it. You have to go in with a base of knowledge at the front and then again at the back. That's how I use it.
Avoiding AI Shiny Object Syndrome
Gianluca Fiorelli: Yes, totally agree. One question, talking about AI, and now I return to the practical side of AI. One of the big problems, positive problems, you can say, is that AI in general, and AI search specifically, is bringing us something new every now and then. Substantially every day we have something new. Maybe during the summer people took it quietly and had their holidays, so it was not so crazy. But ACP, UCP, MCP, all the agentic stuff, then OpenAI model changes, checking out what changed with the model, and so on. There is a risk. Don't you think that sometimes there is the classic shiny object syndrome that already existed before, chasing these new things? It was already present before, but now it's sometimes even stronger.
How do you somehow put a brake on your enthusiasm about AI when it comes to, “Okay, let's do the real work”? How do you decide whether to pursue something announced now or later, since it's not relevant for us right now? And especially, how are you able to explain this sometimes to stakeholders? Because sometimes the problem is that the stakeholders are the ones who are more enthusiastic and want to do something related to this shiny object.
Abby Gleason: I like to go back to user research. I know not everyone has the luxury of a whole user research team, but even just talking to your one customer support person about the key problems people have here can be really helpful.
To your point, I think there's a big shiny object syndrome. A lot of websites I go on recently have this humongous search box in the middle of their homepage that says, “Describe in detail what you want to do.” This is something I get really curious about because I, myself, don't love to search that way on other sites. That's just me personally. But it makes me wonder: was this a shiny object, or did they get this feedback from users who said, “I want to go into detail and describe something and then get this personalized result”?
I'd always go back to what the core three are. I like to always have three big problems that users have that are the friction between them and signing up for your product or buying your product, and make your website and every experience that you put out there, your content, everything, revolve around solving those problems.
Sometimes maybe that can be done with something that's a shiny object. Maybe that does solve their problem, or maybe it doesn't. If you have a really strong base of what your users' key three problems are, then I think you have an answer that you can bring to leadership and at least have some backbone of, “Here's why I would or would not recommend that.” That doesn't mean you always will be able to get away with not testing that out, but at least it gives you some base in reality, in research.
Gianluca Fiorelli: Yes, indeed. As a curiosity, I always thought that in the past, especially if you were putting internal search up front on a website, it was because your website had big problems in making people navigate inside it.
Now, I think that can be tricky and can be interesting for certain types of businesses to translate conversational shopping, which can be, for instance, on ChatGPT or AI Mode, inside your website. People are starting to get used to this kind of conversation, but also as a way to nudge the user to a faster conversion.
Because AI is very convincing. It's very able to sound plausible. When you are talking with your AI about, “This product would be perfect for you because of this, this, this,” it's like once people were saying, “Because Google says so.” Now, “Because AI wrote to me or told me,” or even giving some name to the AI, like Chatty, or something like this.
Maybe there are opportunities, but I'm totally agreeing with you. Before going deep and immediately embracing something, do a pilot, test it, and see if it's really relevant for you.
Why Websites Should Still Leave Room for Discovery
Abby Gleason: Definitely test it. One thing, too, is that I think websites can and should also be used for discovery. When I see these AI boxes, I think maybe it's a good idea, but it's always too early in the funnel for me. It's putting a lot on the user to know what they want. One example is Airbnb or Vrbo. Have you used Airbnb before?
Gianluca Fiorelli: Yes, obviously. I also use it as a good example of what I call the poor information gain architecture because of the way they are categorizing all the apartments.
Abby Gleason: Exactly.
Gianluca Fiorelli: Searching for a house in a tree, a tent in the desert.
Abby Gleason: They have a fascinating search ecosystem. But one thing I think of with them is, if I go on Airbnb and I'm looking, let's go back to my Mexico City vacation, I don't know exactly what I want yet. I can't say for sure that I want all these XYZ features. I need to be inspired and to discover. Part of the reason I like going on my Airbnb app, I actually go even when I don't have a trip, is because it's fun for me to look through all the different experiences and options.
I ended up in this beautiful loft in Roma Norte, but I wouldn't have been able to say that. I needed to explore the different areas first. I needed to explore what types of apartments were available, and only then could I say what I wanted. It's the same with an e-commerce site. I may not know that I want this exact shirt with this exact feature. It's only through looking through the options that I become knowledgeable about the thing that's right for me.
When I think of shiny object syndrome and putting an AI box on everything, maybe you should just think first about what the experience you want people to have and maybe enjoy the search.
Gianluca Fiorelli: This is very often a topic in travel, to create navigation based on experiences. Sometimes it's hard to make it understandable that experiences are not just creating a new taxonomy about things you can do. Experiences are something more intangible. I want to have an adventure travel experience. I want to have a relaxed travel experience. I want to have a travel experience that makes me feel I'm learning the culture of the other country. These are more intangible, not like, “I want to go diving in the Caribbean.” That is a feature. It's not an experience.
Abby Gleason: Not everything should be a blank box.
Why Abandoning Informational Content Is a Mistake
Gianluca Fiorelli: No, and this also makes me want to ask you something, because there is a sort of trend, somehow in certain cases justified, of quitting creating informational content because informational content is cannibalized by AI. I sincerely think it's a stupid idea.
Abby Gleason: I agree.
Gianluca Fiorelli: I'm very honest. I think it's a stupid idea. One, because you are renouncing your influence on what you were saying, the inspirational moment of a search customer journey. Secondly, if we want to talk about AI and LLMs, we know it's really based on knowledge graphs, semantics, expertise, and how you, the brand, relates to a topic. If you are abandoning all the informational phase, you are substantially relegating yourself to maybe being considered an option for a further stage. But you have not conquered that kind of influence we were talking about before in search, in the answer, even if they are not clicking on you. From your expression, I think you are agreeing with me, but I want to hear you.
Abby Gleason: I couldn't agree more. I think there's a distinction with informational content. Ten years ago, I think of writing content for my clients, and you used the word relevance, and that's the key. I think we've all been there where maybe we've written something that's kind of off-topic for our client because you have the potential to drive traffic.
That is the only type of informational content that I think is cannibalized and is over for us, and thankfully so. Whenever someone says, “Content is dead, what's the point?” I just use my own experiments and actual data. There's nothing further from that. Even if you don't get as much traffic as you used to six years ago, you can still be listed as the top citation source for that. So you're still getting impressions, and you're still creating influence.
I'd say it's a short-sighted viewpoint to say, “I'm not going to create informational content because AI Overviews will just cover it.” No, you'll be the source cited. You could still drive traffic, and ultimately you're telling AI search, like ChatGPT, AI Mode, AI Overviews, et cetera, “Hey, I'm an expert in these areas.”
If you stop creating content that shows your expertise and relevant information, eventually, you won't be mentioned at all, and no one will find you.
We're still creating this category FAQ content, and it is just performing like crazy. It's showing up in AI Overviews. It's driving traffic. It's improving rankings and visibility. We're also still making informational content on YouTube and our blog. Are we doing it at the same pace and with the same type of topics we did five, six years ago? No. We're focused a bit more mid-funnel, bottom-funnel, but we're still creating hordes of informational content, and it is absolutely still performing for us. So if we stopped it tomorrow or deleted some of it, we'd see instant declines.
Gianluca Fiorelli: Yes, also because I think that sometimes people forget that every page on a website is a potential landing page. So it could be this blog post that you wrote about a middle-of-the-funnel topic, renting apartments in New York City, for instance. And from there, people could follow up inside the website.
In real estate, we know that the conversion is not something immediate, but something that can take one month or even more than one month. You can track this person somehow. And so you can see if he's returning, not returning, and if he's returning, what he's visiting, et cetera.
In the States, you don't have so many problems with the cookie law, so you can eventually also personalize even further the experience of returning customers. Informational content many times, is also the correct type of content to make people return to your website.
And also because of repurposing, as we were saying, we need to be visible on other platforms where people are searching. That same article can become a TikTok educational video, a story on Instagram, or a LinkedIn stats kind of post, and so on.
Abby Gleason: Absolutely. Content is authority. Think about anyone in SEO: if they never posted anything or shared anything or wrote anything, how would you know that they are an expert? It's like that with any brand. What you put out there and show that you have experience in is how both humans and AI agents understand that you're an authority, and so you have to be consistent.
And that's where just using AI to write all your content can be a threat. I'm not saying don't use AI in your process, but definitely have a human editor in the loop and incorporate proprietary data if you can, because that's another way that you're going to show your influence and authority, not just recycling what's already out there. You have to have some form of new information, ideally, in most of the content that you publish.
Gianluca Fiorelli: Yes. Or you should be a true curator, in the sense of not just repeating data from others but really creating new insights using the data of others.
How Big Brands Can Show First-Hand Experience
Gianluca Fiorelli: And I have a curiosity because we were talking about experience. Experience is something that, in this case, small businesses are able to demonstrate more easily. You were giving the example of a car dealer who is very expert about a specific type and model of car.
But in the case of big enterprises, sometimes showing experience—not expertise, not authority, but direct experience—is more complicated, or it seems to me it can be more complicated because of the anonymous nature of a big brand website. What are, in your opinion, ways to show experience even if you are a big brand, but you normally tend to be a faceless company?
Abby Gleason: Do you have an example of a company? I'm curious to hear.
Gianluca Fiorelli: For instance, what I'm doing for a company in travel is we decided to show experience by using people working in the destinations where they are present. It was quite complicated in the beginning, but then we created essentially a big Excel file, trying to make it very easy for them.
A big Excel file where we're sharing all the tips they recommend about the destination they are responsible for when tourists and clients are coming, for instance, to the Dominican Republic, Peru, or the United States.
So we're using that information and then asking for permission because it's a European company. We need to ask for permission to use the name, the face, et cetera, of this brand destination manager. And we're using them as a tiny bit of content on the landing page of a destination, with something like, “Our brand manager recommends…” “If it's the first time you're traveling here, I recommend…” You have a face, so it's about speaking directly to a visitor. “I recommend you choose this place and not this other place here in the Dominican Republic because it's better, it's more comfortable, and it's less shocking,” whatsoever. Even going a little bit against selling everything. So this is a way that we found for this kind of company. But I was wondering if you had to deal with this kind of situation too.
Abby Gleason: I will admit this is one that I don't know that I have a super good answer on, so I don't want to say something I don't feel confident on.
Gianluca Fiorelli: Because I'm working with every kind of company, it's one of the things that I saw as a difference between a small to medium company and a big enterprise company.
For more on the ideas discussed with Abby, listen to our conversations with:
Emina Demiri-Watson on taking a broader marketing perspective, using audience research to understand customer journeys across channels, and setting priorities around business goals amid AI distractions.
Amanda King on adapting to AI search, moving beyond keyword obsession, and understanding the limits of LLM tracking while keeping SEO grounded in marketing fundamentals, user needs, and measurable business impact.
Winning Stakeholder Support With Small-Scale Tests
Gianluca Fiorelli: But talking about big enterprise companies, with all these years of working in enterprise, what are the tricks you have learned to sell stuff to do to the stakeholders?
Abby Gleason: What's something I've learned to get stakeholders to see my side of things? Mini tests have been a little superpower. Basically, running small-scale experiments, getting some data, and then bringing data to stakeholders. I find that that is what gives me confidence when I come into meetings, sometimes all the way up to the C-suite, if I actually have numbers to back something up rather than just opinions or vibes.
One example was the tool that I was talking to you about, agents. But actually, a better example comes to mind. We had been running a small-scale program to update some of our programmatic pages with FAQs. And so we ran that as a test and basically compared the pages that we updated to the pages that we didn't, and we saw instant gains.
That is actually what allowed me to go to the stakeholders recently and say, “Look at the performance of this small subset. Give me budget for the engineers to basically scale this across…” We ended up scaling it across 400,000 pages rather than four or five hundred. And because I had that small-scale test to say, “Look at the gains here. This is what I think we could achieve if we scaled this up 100K,” I got the green light, and it was a heck of a lot easier than saying, “Just trust me on this one.”
Because I was able to say, “Actually, you don't need to trust me. You can trust the numbers.” So that's my hack: running things as a small-scale experiment, validating it, and then being able to take that to leadership and say, “Hey, give me more budget,” or, “Give me the green light on this project because I can prove to you that I can make you money.”
Gianluca Fiorelli: Yes. Totally agree. And also, it's maybe the best way to not be relegated, especially in companies where product is leading the orchestra, to the end of a sprint. It's a tragic situation because usually delays make you jump to another sprint.
Abby Gleason: Absolutely.
Gianluca Fiorelli: It's quite frustrating sometimes.
Abby Gleason: It can be. You have to speak their language. Product speaks the language of money, and so SEOs need to as well. You can't just say, “Crawl budget is a problem. These indexation issues.” No one's going to hear you. You have to say, “I have an issue that I believe is costing us X dollars.” And then you can say, “You need to fix this indexation issue.” Lead with the actual business problem, and you're going to have a lot easier time getting your resources and getting leadership to pay attention to you. Let's talk dollars.
The AI Search Myth of Reddit as a Shortcut
Gianluca Fiorelli: One last question, and maybe a fun one. SEOs still have many myths. What are the myths of AI search that you are starting to see, and maybe thinking, “Oh, no, again, people are saying this? It's a myth. Why do they believe it?” Are you starting to see some myth emerging related to AI search, at least based on your experience?
Abby Gleason: Some what?
Gianluca Fiorelli: Some myth, like in SEO, keyword density was a myth.
Abby Gleason: I got it.
Gianluca Fiorelli: And now in AI search, I don't know what could be the myth. Like, you have to write very short phrases of 120 characters. This kind of myth is related to chunking.
Abby Gleason: I knew you were going to say chunking. Chunking comes up, but I don't know if this is a myth. I find this whole Reddit strategy thing kind of insane and laughable. Basically, I feel like I keep seeing agencies or freelancers who say, “I can get you featured in all these Reddit threads.” Reddit is allergic to anything that's disingenuous. You're going to get banned. You're going to get absolutely berated. I think it'll actually do your brand a lot more harm than good.
There are really only two ways to do Reddit for AIO. The thing I keep seeing is, if you want to be featured in AI Overviews and Chat, you need to be on Reddit. I agree, but you need to get there through reasonable means, and there are only two ways to do that.
One is going back to having a good product. Literally just have a good product, and people will mention you. Going back to Mexico City, when I'm looking up the best restaurants to go to, the restaurants people were recommending were unanimously decided as good by people. Then you see the upvotes, and you're like, “Okay, 700 people agree that this place has the best tuna tostada, and so I have to go there.”
And then the other is, I do think that there is a way that you can put forth your brand, but you have to be incredibly authentic and transparent about it. I actually had a case where I was researching on Reddit, and a founder was like, “Hey, I built this small app. You can use it for free. No worries if it's not helpful.”
That was the kind of thing where I was like, “Okay, cool.” I don't feel that this is a problem because they're trying to offer a solution, and they're being very realistic that they're the founder and they're trying to be helpful. If it were someone who created a fake account and was just spamming this brand name and, “Everyone loves this,” it's just going to be really apparent that that's not helpful.
Gianluca Fiorelli: It wasn't even working before.
Abby Gleason: No. And so it's not working now. I think the myth is, if featured on Reddit, then we'll win at AI search. No, you have to be featured in the right way. And I think the best way—
Gianluca Fiorelli: And then also, it's like making Reddit something as if there were just a couple of hundred subreddits. There are so many subreddits about everything and hundreds of subreddits about some specific topic. You really need to know well and investigate well what is the subreddit that really matters to you, and do audience research in the subreddit ecosystem in order to understand how you can eventually interact with Redditors.
But I'm not so used to Reddit because it never was so popular here. At least in Spain and Italy, we had very specific forums, which are very local, and even "worse" in terms of transgression and severity of members of a forum when a brand is coming.
Abby’s Favorite Memories From Mexico City
Gianluca Fiorelli: One hour already. Really fast. Just one question, and it's not about AI, it's not about work. You talk a lot about Mexico. What were the things that you loved the most about Mexico City? Are you just saying Mexico City?
Abby Gleason: Yes. I went there last year with my sister on vacation for a little less than a week. My favorite thing about Mexico City is a tie between the food and—I have to look it up. Give me one second. It's Lucha Libre. For $3, three American dollars, we got to go to a Lucha Libre show, and we had the most fun. I have not laughed that hard in a long time.
Gianluca Fiorelli: It's quite wild.
Abby Gleason: Really? Have you done it? Have you been to Mexico City?
Gianluca Fiorelli: Yes, I visited Mexico City. I never went to a Lucha Libre fight, but I know it. I know it very well.
Abby Gleason: You have to go. And a hot tip: if anyone is thinking about going to Mexico City: do not go online. They'll charge you $60, $70. Just go straight up, and it's $3. At least it was a couple of years ago.
That, and then the other favorite thing we did—I keep mentioning Airbnb, but we did one of those experiences and worked with Chef Raja. We learned how to make authentic tacos al pastor, and I still make that recipe for friends.
It was an absolute joy and highlight. And just in general, the people. The people were my favorite part. We just met some wonderful, wonderful humans. I'd go back in a heartbeat.
Gianluca Fiorelli: Yes. All Mexican people, all our friends of Latin America, are really wonderful people. And also Spanish ones. European Spanish are great people too.
Abby Gleason: I have to get to Spain. That's huge on my list. I love to eat, and I love tapas and wine. I've heard great things.
Gianluca Fiorelli: So we can make a search session live streaming in the same room.
Abby Gleason: Perfect. I love that. I will take you up on that.
Gianluca Fiorelli: Thank you, Abby.
Abby Gleason: Thank you so much.
Gianluca Fiorelli: Thank you for being my guest and the guest of Advanced Web Ranking today. And thank you all too, friends. Remember to subscribe to the channel and ring the bell so you will be notified when a new episode like this wonderful one with Abby is coming out. Thank you and bye-bye.
Podcast Host
Gianluca Fiorelli
With almost 20 years of experience in web marketing, Gianluca Fiorelli is a Strategic and International SEO Consultant who helps businesses improve their visibility and performance on organic search. Gianluca collaborated with clients from various industries and regions, such as Glassdoor, Idealista, Rastreator.com, Outsystems, Chess.com, SIXT Ride, Vegetables by Bayer, Visit California, Gamepix, James Edition and many others.
A very active member of the SEO community, Gianluca daily shares his insights and best practices on SEO, content, Search marketing strategy and the evolution of Search on social media channels such as X, Bluesky and LinkedIn and through the blog on his website: IloveSEO.net.
