AI Broke the SERP: Here’s How Enterprise Teams Can Still Win with HubSpot

Tonia Speir (00:14)
Hey everyone, and welcome to SmartBug on tap. I'm your host, Tonia Speir Senior Enterprise Account Executive here at SmartBug. Before we dive in though, don't forget to subscribe so you never miss an episode, especially if you're leading marketing rev ops or content strategy at the enterprise level and want to stay ahead of where HubSpot, AI and SEO are heading. Today, we're going to be digging into a major shift in digital visibility, answer engine optimization or AEO. 

With AI-driven search results becoming the norm, whether it's Google's Search Generative Experience, SGE, Bing Copilot, or AI sumMaris in chat, traditional SEO is no longer enough. Enterprise brands now need to optimize for how engines answer, not just how they rank. Joining me today is Mari Boni, SEO specialist who's been helping enterprise teams navigate this evolution and rework their HubSpot content strategies to stay competitive. Mari, welcome to the show.

Want to kick us off with a quick intro?

Mari (01:12)
Hey, thank you for having me here. Yeah, sure. So I've been working in SEO for a few years now. It's been five years or so. And before that, I used to be a quality writer. So I was analyzing search results pages for Google. So I'd say it's been almost seven years that I know a little bit about search engines and everything.

I'm really glad to be here because it's true that it keeps evolving all the time and that's the reason why I'm glad we can talk about this.

Tonia Speir (01:46)
That's awesome. Well, thanks, Mari. You say you know a little bit. I think you know a lot, especially compared to me. So excited to have you here. We're calling this episode AI Broke the SERP and for good reason. The way people discover content online is fundamentally changing. It's no longer about climbing to the top of Google's results pages. It's about the answers that consumers are getting in these AI tools and what they choose to surface.

Mari (01:50)
Meh.

We'll try.

Tonia Speir (02:13)
that shift has had huge implications, especially for enterprise teams with a lot of content. They have complex customer journeys and they have a lot of stakeholders that are a part of the buying process. So Mari, to get us started, can you break down what AEO really means and how we got to this point?

Mari (02:30)
yeah, sure. AEO, it stands for Answer Engine Optimization, as you've said before. For many, it's the idea of structuring content to answer questions directly for AI tools or voice assistants. It's often presented as the next evolution of the SEO. And we have many other acronyms for this. We have GEO.

AIO, LLMO, so many. So about how we got to this, that's a good question, but I think it was inevitable with the rising of AI. We had the chat GPT that came out in 2022, but I think it got worse for us with the AI overviews last year in May 2024. And now in April, we've had the launch of AI mode from...

Google. yeah, that's a

Tonia Speir (03:27)
Well, Mari, thanks for telling us a little bit about the background. I know AEO, it's a mouthful, it's hard to say, was new to me. So tell me, how are search experiences changing for buyers?

Mari (03:27)
release.

I think it's become more emotional, more tailored to everyone's personality. Since you can develop a relationship with a chatbot now, you deliver way more information about yourself than we used to in the past. And that allows the bot to adapt the answers to you as a specific buyer. So for sure, it has a lot of impact on what we used to do in SEO.

I'm not saying it was not the case with the search engines though, because it already was in a way. I don't know if you're familiar with Google patents, but I don't want to be too much tech today. But these are legal documents that protect Google, their inventions. And there's one that came out last in 2023. It was called Query Composition System.

And it clearly showed that Google is already using a lot of things to tailor our experiences in SERPs, just in case SERPs, it's search engine result pages. Prefer to be clear with every acronym that we're going to talk about. There are so many. So yeah, they were used to use all of this. So that's our history brother, our location, our preferences.

Tonia Speir (04:51)
There's a lot, right?

Mari (05:02)
The time when we were asking for something, of this were analyzed. And it was to better understand the context to help you ask the right thing to example, if you are close to a movie theater, Google could know and provide predictive queries around this topic to help you ask the right thing.

But I wouldn't say it was as tailored as today with AI chatbots, because the way we express ourselves when we ask questions, we give so much about ourselves. And everything is stored in AI's memory, let's say. What has changed? I would also say it has kind of narrowed down a lot more the choice you have as a buyer, I think. When we were able to stand out

more easily in the past with a very good organic strategy. I kind of feel that nowadays with the large language models, if you don't have a good brand authority, if you don't have a good reputation over the web, it's going to be more difficult to appear in buyers' eyes. And answers have become, at least to me, very much all the same. When you check AI overviews, there is always a pattern in the way it's answered.

they're everywhere and we have to adapt.

Tonia Speir (06:19)
And I heard you throw another, you used the long form of another common acronym that people are throwing out, is LLM. And so that is really important for when we talk about what's the risk for enterprise brands that have been relying on legacy SEO tactics. And now how do they instead shift and train for those LLMs?

Mari (06:41)
it's a good remark because it depends on what you actually mean by legacy SEO tactics. Because for sure, if you were referring to keyword stuffing like we used to do back in the days, content optimized only for bots and not for people.

But technique, for sure, enterprises that do that can worry. That's true. Google is getting sick of this, and it's been years. If you get a look at the recent Google core updates, only the last two years, it's enough to see that they have refined their algorithm throughout the years. And they really want us to focus on good content that is optimized specifically for users.

So that's why we're not talking about that SEO. It's not the same. And I think that's the reason why you're asking this question. There are so many people who think that SEO is only about keyword stuffing and backlinks and that's it. But it's so much more than this. that's what makes it so fun for us, for people who work in SEO today. I don't know if I have the right to explain.

What I think is SEO in order to advertise my discipline. But it's just that it's not, it's not about keywords. It's about understanding what's behind a user's query and to provide the adequate satisfaction to this query organically. It's about organically satisfying a user. And this implies so many things. And these things are really important for AI optimization as well.

Tonia Speir (08:16)
I think I hear that a lot, like from questions they're like, we have good SEO, you know, why should we worry about this? Or, you know, is there still a place for both in a strategy? Do you still need to do traditional or legacy SEO tactics along with this new, you know, strategy of training for AI? Do they both have a seat at the table?

Mari (08:38)
yeah, yeah for sure. I think it's because actually AI is not something that different from SEO. think AI is not killing SEO, it's exposing weak SEO. I don't want to sound mean when I say that, but it's just that I don't worry for enterprises that have worked on their SEO the right way actually. And I think that...

People who are panicking over AI maybe never understood what is SEO, the real one. Yeah, the game has changed, it's true. We don't have the same tools no more, but the goal remains the same. We just want to solve users' problems and we want to do it better than anyone else.

whether it's YouTube, whether it's AI models, whether it's search engines SEO still applies.

Tonia Speir (09:28)
So why does it hit enterprises harder than SMBs?

Mari (09:32)
I don't think it's really about the size, I would say it's more about the strategy. Because right now SMBs can struggle actually because they often like brand authority. They like backlinks, reputation, and now that authority matters way more than ever with LLMs, enterprises have this advantage at least.

But it's true that some SMBs can outperform enterprises because they've been intentional with their SEO and the content strategy. I think the reason why enterprises are not immune and can struggle is actually because, and that's a good thing, what I'm going to say, they invest a lot in content, in their content strategy, which is a really good thing.

but they don't align it with the user's needs. And it becomes writing at scale, writing all the time, a lot of content, but it's not relevant. And that's a big issue for either SEO or AEO. I'm going to use the acronym. So yeah, when you're producing at volume, you have better chances of producing low quality content. And that does not align.

with good practices of today. And I think there might be something else. It could be a question of niche as well, because I don't know if you know your Money Your Life websites. It's when it's about health, finance, legal. All of these niches are more at risk because

There is a stricter scrutiny. Both Google and large language models in general are very cautious. They don't want to give untrustworthy information. So that's the reason why niche are sometimes at risk.

Tonia Speir (11:23)
How do they police

Mari (11:26)
related, I think, to authority. So the higher your reputation is, the best it is for them. There is something when it comes to AI that's called hallucination risk. they have so many occasions to say the wrong thing because it's not intelligence. It's just a tool.

And that's the reason why they try to go to sources that they know are trustworthy. And there is many different factors. To be honest, I don't have the pretension to know them all yet. We are still investigating all of this. But I think it's the same as it was with the search engines, the classical ones. brand authority with the backlinks.

Quotes, citations, UGC content, it's user-generated content, so reviews over the web, stuff like this. The more your brand is spoken of, the best it is for them.

Tonia Speir (12:25)
Okay, so that's some of the things that improve than their perceived authority. Okay, so now that we know that the key takeaway there is that visibility now means being summarized, not just being clicked, and enterprise content teams have to adapt to this ⁓ quickly. We know it's all moving faster than I think we anticipated. how do you, can you explain just in, for those of us that aren't as close to it,

How is AEO different from traditional SEO?

Mari (12:54)
This question, it just makes me smile because it's a huge debate among the SEO community so far. There are many people who think it's exactly the same. There are many people who think it's completely unrelated. I'll try to have a balanced discourse, balanced talk and say it's a little bit in between the two. It's different and it's not different.

It's just like SXO back in the days, like five or seven years ago, we were using this new acronym and it stands for search experience optimization. It's when people discovered that UX matters in SEO, which is true. But we didn't stop doing SEO. We just added a layer to it. Same when I mentioned SEO for YouTube earlier.

We're not inventing a new term. We just do SU on YouTube and it's going to be the same here. It's part of the SEO evolution. And the reason why I want to specify this is because I wouldn't like people to get into FOMO, fear of missing out and think, my God, this is crazy. This is completely different. We have to adapt, change everything. No, no, no, please don't change your strategy. It's a good thing.

Tonia Speir (14:09)
Don't throw the baby

out with the bath water as in Southern Tours.

Mari (14:10)
my, yes,

exactly. That's exactly what I mean to me. Again, it's my perception. So I don't want to speak for everyone else. I think AEO is mostly a rebranding of what good SEO already was. I think we'll have the opportunity to dig into that further, but there is one thing that is interesting to highlight though. LLMs and generative AI.

These are models that work with what is called RAG. I think in English it's retrieval augmented generation. That means they retrieve data from search engines like Google and Bing and everything. search engines, these are our first targets in SEO. because we operate on search engines, so it's kind of intertwined in a way. And Google themselves confirmed this.

because right now they've got an antitrust trial with the US. Poor of them. But internal emails leaked out and we have access to them. And one of them showed that Gemini actually uses a search grounding. It means that they basically check the SERPs before answering. And I'm pretty sure that they're not checking the bottom of the SERP. It's so the higher ranking SERPs.

the higher the chances you're going to have to appear and to be shown into AI tools like Gemini, for example. there are nuances to optimize for large language models. That's true. And I think the differences lie in the way they retrieve data when they process a query from a user and then

redistribute the data.

the query-findout process, what it is, it's that models that are powered by AI aren't just matching your page to real questions. What it does, they are imagining every possible way someone could ask for what you offer.

these possible ways, we call them, it was called synthetic queries. It's latent queries actually. They want to provide the best possible answer to someone by anticipating everything and be fully aware of the user's context. And that's something we really need to know even more now than before, because we were actually doing this already in SEO

But we really want to capture latent intent, anticipate the needs of someone, because that's how it works now.

So that's what is query final

Tonia Speir (16:39)
So Mari, there's obviously nuances for how you optimize for the LLMs based on how you've explained how they redistribute information. So what factors influence whether content is included in an AI-generated answer or not?

Mari (16:53)
It's a very good question, a very frustrating one because I would like to say so many things. And again, I just would like to make a little disclaimer. We are still all investigating in the SEO community, so I'm going to say things. But we don't have the magical recipe to appear in the AI-generated answers yet. We try. And we have.

We already know a lot, but we don't know everything. And actually, it starts with the basics that we do in SEO as well. So first things first, indexation. Make sure that your page is crawled and indexed because a page can be crawled but not indexed. That's not the same thing. And if your content sucks or is thin, maybe you will never be indexed. So just make sure of this.

There is also the technical aspects. Today, AI models struggle parsing JavaScript, so make sure that it's rendered in HTML. Again, in SEO, we usually advise for this as well. And then it's about content, and there are many things to say about this. The first one is EEAT, another acronym.

It stands for experience, expertise, authoritativeness and trustworthiness. I made it. These are very important and they've been important for years, actually. And even when I was a quality rater, back in the days, it was only EAT. Now we have two E's and they matter a lot, especially when it comes to your money, your life websites. Again, the goal of these, sorry, is to...

show that you are an expert at what you do. So the best way to do that is to provide author bios after you've written a blog post, quotes, external validation. Just make sure that you provide the best content ever and that it's clearly obvious to everyone by the way you've written it and by the way it's...

supported

Tonia Speir (19:01)
And quick question for you here, does timeliness matter? Because I know you said that, you know, they won't search to the bottom of the, to the bottom, right? But one of the things that we used to traditionally do for SEO was to reoptimize historical content optimization where we would republish or, you know, so it was fresh new, it's still maybe evergreen content, but putting a new spin on it.

Mari (19:03)
Yeah.

Exactly.

Tonia Speir (19:24)
Is that still important? Like even if people have good SEO results today from older content that they put out there, what about historical content optimization?

Mari (19:35)
That's a very good question. When I said bottom, I was mentioning the bottom of the results because it's true you can have in the SERP old articles that rank really well and you're right. It's true that we really, really need to optimize content regularly, have evergreen content as you said. And yes, it's clearly important. I would say that...

Tonia Speir (19:42)
Mm-hmm.

Mari (20:01)
When you have an article that ranks really well and it's been the case for ages, sometimes I don't touch them because I don't want to take the risk, but it's okay just to refresh the dates and it's okay if you don't optimize everything. As long as the content itself is relevant and up to date, you're fine, but you're absolutely right.

if it's recent, it's better.

is something else that is very important when you want to optimize your content with AI now. It's semantic clarity and reformulation in order to have the semantic similarity. It goes with what we were saying earlier. The fact that LLMs generate synthetic.

queries in order to understand what the user is saying. So you want to cover a lot in your contents. Use precise vocabulary. rich semantic fields, structure your content clearly. It was already the case with SEO again. H1, H2, H3. This is very useful for the bots to understand the order and the...

hierarchy of your content, the structure. And again, short paragraphs. Really go straight to the point. We don't have time to talk about redundant stuff or non-relevant stuff. And it was, again, what we were doing the past years when we wanted to rank in positions zero. When you type something into Google and you have a big result.

and you can click on it if you want to have the whole blog post. But usually it was possible to rank easily with this. Go straight forward, provide the best answer to the user. yeah, talk to

Tonia Speir (21:46)
When you talk about

precise vocabulary, can you give me an example of like maybe a client you were taking through this optimization process and a before and an after type example?

Mari (22:02)
Yes, I think I have many different ones and I'm gonna think of one that's relevant.

There is this client that I'm auditing right now. They provide formations for professionals. And when I was reading the content on the website, I couldn't understand anything. It was too corporates talk.

they were not addressing to the audience. And that's exactly what I mean by this. You provide courses for people, say you provide courses for people. We don't have time for anything else. Just be clear. Yeah, exactly.

Tonia Speir (22:37)
corporate jargon.

it sounds like you might have, you now have a dual strategic purpose, right? Like where on one hand you have to write for optimization against competitors and keywords who are maybe using some of those more corporate jargony type terms. But at the same time, you have to write for the everyday consumer that doesn't think in the way that you explain things and in a more simple way. I once had a creative director, it reminds me like,

Just this example and listening to you talk about it, I had written a strategic plan for a client and she threw it back at me. She's like, what does any of this even mean? And I'm like, well, how could she not know what this means? Like, this is what we do every day. And she's like, no, explain it to me. And remember that our customers don't do what we do and they don't think to ask the way that we ask things. So you have to explain it in the way that they're asking.

Mari (23:14)
Mm, good.

Tonia Speir (23:32)
Just kind of listening to you, explain this example, it sounds like it's a two-fold problem. For one, you still have to optimize for those big jargony corporate words that are industry specific, but at the same time, also write your content in a way that answers the question at the level the customer's asking.

Mari (23:50)
True, you're absolutely right. And I have the same issue with a client of mine because they work in a, it's an automation company. They do automated test systems and it's honestly extremely difficult to understand what they do. So the way we articulate our content, we have to, as you say, we have to address those terms because they are important. And even for the...

Even for AI models and search engines, it's important to mention them because it shows that we are experts at what we do, that we cover this semantic field. And in the meantime, you have to tailor your content and make it accessible to your users. So yeah, you have to be balanced.

And yeah, we mentioned it earlier, but authority is extremely important. The stronger your domain level authority is, the more likely your content is to surface. So that's, it was already the case as well, but even more now.

Tonia Speir (24:54)
Mari, let's talk for a minute about content format and structure. What does AI want? What do they prefer? What does it prefer?

Mari (25:04)
I would say content that goes straight to the point that's really clear. So you need depth. Think about providing definitions right away at the beginning of your content. Give comparisons, FAQs, real life use cases. Basically content that is completely satisfying user intent and anticipating the needs of...

of a user. And as to formats and structures, please, this is a message for the developers. But even for anyone that owns a website, it's really important to use the headings the right way, H1, H2, H3, because all the time when we get a new client, we are very surprised to see that the way that they are used is for style.

just to have big titles and little titles. But that's not only the use of a heading, it's really to show to the bots, OK, that's the menu of your page and that the subtopics of the page here and here, that's where you insert your keywords, where you ask the relevant questions. So very important to structure the content with the headings.

Yeah, I do need to mention for AI, something that is really interesting to dig as a kind of content type, I would say to work more in your industry pages because they know more about the users now. So let's say you are a social media manager.

and you're going to say that to chat, GPT, Gemini, cloud, whatever. You're going to provide many information about yourself and you're going to say, I'm looking for a media asset management. Can you help me with that? And if you are that provides media asset management services, it's interesting to have a page that is dedicated to that kind of audience because you're going to adapt.

what you say to them. If you target social media managers, it makes sense to mention what they are used to, their pain points, what they are struggling with in their everyday life, because the bots will be able to be, okay, this is mentioned over there, so maybe it's going to be relevant to propose this content to them. And about page...

The about page of a company is extremely important because when people want to know more about a company, bots are going to go there to find information. So make sure that it's optimized as well.

Tonia Speir (27:49)
Got it. So we're gonna shift gears here for a second because majority of our clients and those who are probably tuning in are on HubSpot. So I wanna talk about building an enterprise, AEO strategy in HubSpot. How can HubSpot user structure their content for AEO? Think of things like pillar pages, blog formatting, featured snippets. What should they be mindful of?

Mari (28:14)
Very good question once again. What I like with the HubSpot is that they really think with a clustered content strategy. yeah, pillar pages are really important. Make sure that you've got one comprehensive pillar page that targets a broad topic and then make sure you have multiple support blogs answering specific intent-based questions.

In order to optimize those pillar pages, it's the same answer right away. Summarize your content at the beginning. It's easier to digest and to know what's going to come after. Again, anticipate latent queries. That can be done with People Also Ask. When you type a query into Google, you always have a list of questions with People Also Ask. If it's relevant to add it into your content, do so.

because like this you're anticipating other queries that might come up. So yeah, make sure you cover everything. Don't sound redundant. And featured snippets, it's a good idea too, like having tables, lists, definitions.

But I just wanted to add that.

It's important in the content cluster strategy to have the pillar pages and many blog posts, but it takes a specialist to know when you will need to write one piece instead of two, or two instead of one, for example.

Tonia Speir (29:39)
Mari, so looking ahead, AEO, AI, and the Enterprise Funnel, will AEO lead to zero-click journeys, and how should Enterprise teams respond to this?

Mari (29:52)
yeah, it will and it does already. But let's be positive and see the opportunity here because the clicks you'll get have better chances to be of high quality, which is a good thing. But yeah, to be honest, at some point we'll have to change our key metrics. Clicks won't be enough. Impressions are already not enough. So how to respond? Let's say to be everywhere. Be present in bound marketing.

is and has always been the solution in my humble opinion.

Tonia Speir (30:21)
It's interesting that you say that KPIs and metrics are going to have to change, and what should they start measuring? So that's a great lead into what happens to attribution and ROI measurement in this AI first world.

Mari (30:37)
Honestly, that's a very good question and it's moving so fast right now. I can't provide a definitive answer. So far, traffic coming from AI mode from Google is not counted everywhere. It's been confirmed only last week, I think, that the no-referrer bug that prevented data to show has been fixed.

So now you should be able to find the data, at least in Google Analytics 4, as regular search traffic. But it's not segmented, so you can only assume. And I think this is done on purpose, because Google knows deep down how much it affects our website's visibility. So they're being tricky with this one. In Search Console.

there's no way to track the data yet. And this is the biggest issue to me because it's such a powerful tool and it's very useful. But if people are curious, there is a guy, well, a sir, Patrick Stock, he works at Ahrefs and he's done a really nice article and he provides, he talks about all AI models and he says, which ones...

have the no-referrer attribute and which ones do not have it in order to help people track the data.

Tonia Speir (31:59)
Do you think that's coming from Google?

Mari (32:02)
the fact that we can track.

Yeah, well, there is something, the problem with Google is that it was fair play at the beginning. They are an index and thanks to us, because we grant the content, they are able to sustain, if I can say that. But I feel that now they've broken the pact because they are taking the content and they're really

redistributing for themselves and they don't even quote the websites that they've used. So yeah, that's not fair. But the community is rising. So at some point I think it's going to help. Well, I hope.

Tonia Speir (32:40)
Interesting.

Well, tell me what should, so our clients, prospects that are tuning in with us, like if you have words of advice for their teams, what should they stop doing now? And what's one thing they should start doing this next quarter?

Mari (33:02)
Oof. Okay. I don't want to sound like the annoying mama over there, but there are recommendations that I would like to give. Whether you're a member of an internal team, a business owner, even a user. Because when we say that AI broke the SERP, it's the case because it disrupts our search habits.

but also because AI is getting everywhere and it has changed everything even the way we write. yeah, remember that even though it's called artificial intelligence, it's not intelligent at all. Don't take everything for granted.

I've had issues way too often when I was writing content for clients and I got wrong answers and I knew because we talk a lot with our clients. But it's a shame because so many people ask stuff to those tools and they don't verify the sources. So it's an issue regarding information, but it's also an issue regarding expertise.

How can you make sure that ChatGPT or Gemini told you the right stuff if you don't even know what you asked in the first place? So be careful with this. Like really, yeah, remind yourself that you are talking to a database and it only retrieves content that's been stolen for existing websites, existing original content. This is very important to keep in mind because...

AI operates on a paradox right now. It gives away content that is now being written by AI. So you're fitting AI with AI and at some point it's, we're going to have a problem with that. And we already, we already have one. So yeah, as always, the goal when you're right should be to stand out, not to take the easy road, I would say. So yeah, as a company, if you want to be a thought leader,

You can't allow yourself to write with the AI. You need to be different. You need to be better, I guess we could say. I'll only say another one because I have so many things to say regarding AI. But just remember as well that it's not legitimate. AI is not legitimate at all. We are still waiting for governments to legislate the tool. Be careful when you add.

private and sensitive data because you don't know where it's going to end up yet. For example, I would never add anything related to our client database at SmartBug because this information is very sensitive. We don't know where it's going to end up when we write to generative AI tools. So it's important to be careful with

Tonia Speir (35:42)
Your story, Mari, reminds

me of Dr. Seuss's book, The Sneetches, when they produced the machine to put the stars on their bellies. And then you couldn't tell who was from what side of town because they all looked alike and they had no differentiation. So when you talk about AI, talking to AI and being created by AI, it just reminds me of that Dr. Seuss story of not being able to distinguish once it's out there.

I think the key takeaway there is that AEO isn't a trend. It's the next evolution of digital visibility and the faster enterprise teams can move in this direction. The more propensity they're going to have to own and win trust in Mindshare.

So such a helpful breakdown of what's really changing in the world of search. Mari, thank you for helping us as generative AI keeps evolving, enterprise teams have an opportunity here to lead if they align their content strategy with how engines answer and not just how they rank. ⁓ Thank you so much for joining me today and the thoughtful conversation. We really appreciate you being here. And for those of you tuning in, if you want to dive in deeper to AEO, ⁓

Get an audit of your HubSpot strategy or just ask a few follow-up questions. Both my email and Mari's email are in the episode description. We're always happy to connect and always up for a lively conversation. Make sure you hit subscribe so you never miss an episode of Smart Bug on Tap and we'll see you next time.