
Google AI Mode is reshaping rankings beyond the traditional SEO norms with a new focus on understanding complex queries that move beyond keywords and rely on broader signals from the user. So how do you optimize for this?
Find out how to optimize for Google AI Mode in this episode of the Majestic SEO Podcast with our fantastic panel of guests: Araminta Robertson, Charlie Marchant, Nikki Halliwell, Pam Aungst Cronin and Rana Abu Quba Chamsi, PhD.
Watch On-Demand
Listen to the Podcast
Transcript
David Bain
Hello and welcome to the February 2026 edition of the Majestic SEO panel on how to optimise for Google AI mode. I’m your host, David Bain, and joining me today are five wonderful guests.
So let’s find out who they are. Starting off with Nikki Halliwell.
Nikki Halliwell
Hi, I’m Nikki Halliwell. I’m a Senior Technical SEO consultant at Journey Further. I also have my own freelance business and a weekly newsletter called Tech SEO tips.
David Bain
Thank you, Nikki. And also with us today is Charlie Marchant.
Charlie Marchant
Hi, I’m Charlie Marchant, and I’m the CEO at Exposure Ninja. We’re a marketing agency specialising in SEO and AI search. You can find me on LinkedIn.
I spend a lot of time trying to make very dense, technical AI search topics much easier and more accessible for business owners and others with a commercial focus.
David Bain
Thank you, Charlie. And also with us. Today is Rana Abu Quba Chamsi.
Rana Abu Quba Chamsi
Hi. This is Rana Abu Quba Chamsi. I’m based in Switzerland, and I am a consultant and trainer in the domain of SEO and AI. You can find me very active on LinkedIn, so you can follow me.
I specialized in small businesses and startups and associations. So if you need any advice, follow me on LinkedIn, that would be a pleasure.
David Bain
Thank you, Rana. And also with us today is Araminta Robertson.
Araminta Robertson
Thanks, David. Hi everyone. My name is Araminta Robertson, and I am the Managing Director and Founder at Mint Studios. We are a content marketing agency that specialize in the FinTech and financial services industry.
Follow me on LinkedIn to find out tips on everything content marketing and also some stuff I try and be useful and helpful whenever possible.
David Bain
Thank you. Araminta, and also with us today is Pam Aungst Cronin.
Pam Aungst Cronin
Hi. I’m Pam Aungst Cronin of Pam Ann Marketing and Stealth Search and Analytics. Both focus on SEO, PPC, analytics and AI. One, we work directly with clients. The other, we do on a white-label basis through other agencies, and you can find me on all the socials at pamannmarketing. I’d love to connect.
David Bain
Thank you, Pam. So five great panelists there. We’re gonna have a great conversation, an up-to-date conversation, what’s relevant now, how to optimise for Google AI mode.
We have the panelists share some suggested questions beforehand, but if you’re watching live, it’d be great if you have some questions or some thoughts on what’s being discussed, then please feel free to share them in the chat, and we’ll try and incorporate that in the discussion as well.
But let’s start off with Araminta. You suggested beforehand that we talk about the types of queries that are generally used in Google AI mode at the moment. So how would you summarise that?
Araminta Robertson
AI mode, at least in Europe, is just one tab, right? Your default experience is still the regular 10 blue links. But I think most of us will agree that eventually the plan is for AI mode to be the default.
I think what will probably happen in the way Google’s going to route different types of queries, is the AI mode, or you could argue Gemini, in a way, will be used for those more complex queries, like how to perform an SEO audit, or where to find the best restaurants for XYZ, but for maybe more commercial intent or navigational intent queries like “Majestic”, someone who literally just types that brand name and they might stick to the typical ten blue links.
That’s quite an interesting way Google will try to figure out how to use both those ten blue links and the AI mode for more complex queries.
David Bain
Okay, and you mentioned Gemini there as well. So it’s essentially the same thing?
Araminta Robertson
AI mode is like a Gemini 2.0. It’s the same model, but a few differences. It’s basically the same.
David Bain
You talked about being in Europe as well. Today we’ve got Nikki and Charlie, who are based in the UK, Pam in the US, and Rana in Switzerland. I presume you’re seeing the same thing in terms of what AI mode looks like at the moment?
Rana Abu Quba Chamsi
Actually, yes. When it started, we were in a delay, so we had to use a VPN in order to see what’s going on all around the world. But now we are also seeing the AI mode, like what you see in Europe or in the UK, because they have expanded a little bit.
Maybe there are some differences, but actually, yes, the AI mode is number one, if we can say that, in Switzerland as well.
David Bain
And what about in the States, Pam?
Pam Aungst Cronin
Thankfully, we are getting everything as soon as it comes out, and thus far, we’re allowed to use it all. I know there’s a lot happening with different countries, blocking different models and stuff.
I forget which country it is, but someone just blocked or made Grok illegal. Here, we’re trying to do something about DeepSeek, but at the moment, it’s a free-for-all.
David Bain
Okay, so obviously different regions around the world have different access to different results, but we’re all heading in the same direction.
Charlie, you suggested beforehand talking about factors influencing what content AI mode cites. Nikki also suggested something similar as well there. Charlie, shall we start off with your thoughts on that one?
Charlie Marchant
So we work mainly with businesses who are prioritizing how they actually get more leads, more conversions, and ultimately more sales and revenue for the business. At the end of the day, the majority of for-profit businesses are focused on the bottom line.
The concern about AI mode is, if Google search results move more towards AI mode, how are we showing up in those results? Not just as something that gets cited in the side panel, but if someone starts an informational journey, ends up in an AI overview, navigates through to AI mode.
This is the current average kind of journey we’d be looking at, then has a conversation that moves from informational into commercial. Ultimately, the brand wants to be showing up there.
So the influencing factors that we’ve seen, there’s lots of different things going on in the background, but when AI mode first went live, actually, we’d been working with a client who sells iPad Air cases trying to optimize their site to show up more in ChatGPT and other LLMs, other AI chat bots, and we found that a lot of the PR work we’d been doing, the citations we’d been building for them, were referenced in AI mode as well. So they were already visible when AI mode first launched here in the UK, which is amazing, because that’s a great result from other work that had been being done.
In that example, which was “best iPad Air cases”, it’s fairly timely when Apple decides they actually want to release a new iPad. Then, of course, everyone needs an iPad Air case for their new model. Articles that we’ve been publishing on Wired and other tech publishers were being referenced as the sources in AI mode, which means that actually, once we got to a commercial query, something like, okay, which iPad Air case would you recommend me? What’s the best iPad Air case? Something similar to that, we were then seeing our client actually ranking really high in those results, or, I should say, highly visible in those types of results.
So for me, a lot of what’s going on is not only the content on our own site, but thinking about where we’re being referenced, how a product is or a service is being spoken about, and then LLMs and AI mode being able to see the context around that to give users a really good answer when they’re actually getting to the commercial point of their journey.
David Bain
Nikki, what would you like to add to that?
Nikki Halliwell
I’d agree. I definitely think it’s a big factor with backlinks and LLMs always looking for real time verification and ways that they can back up what they’re seeing. And when it comes to content on our own sites, a lot of it is EEAT signals and making sure that instead of just saying, “studies show” if we’re looking at the anchor text, we say, the “2026 Majestic study suggested that XYZ”. These sort of factors are building in and giving that verification and adding into those EEAT signals.
I think another part of it is just the semantic completeness of your answer. It’s not just about where we’d all be trying to get our clients to number one in those 10 blue links. Now it’s more about making sure that we have a complete answer whatever that query is, as long as it fully answers the query, the word count doesn’t really matter. I mean, it’s never mattered really, has it, but, you know, it’s always a metric that clients like to focus on.
What we’ve seen is, even if we’re ranking in position seven organically, that doesn’t mean that you’re not going to show up in Google AI mode or the AI overview. You’re still going to show up in that if you have a more complete answer than whatever website is in the traditional, organic number one spot.
David Bain
Pam, you were nodding away there when Nikki was talking about EEAT and the importance of that in relation to ranking in AI mode. Why is that?
Pam Aungst Cronin
Yeah, so that new E (Experience) has been a bit more problematic than I thought it would be when I really started to dig into how exactly do these machines recognize it. Gemini in particular, because Gemini powers the AI overviews at the top, and it powers AI mode, and it searches the web itself when people are in it.
It’s the Science of Cognitive Linguistics, which I cannot claim to be any kind of an expert on, but it was fascinating, and it made me want our clients to really work harder on demonstrating that E, which is really the only thing that AI-generated content can’t mimic. It can’t mimic your real-world, firsthand human experience that you had on this planet as a specific person at a specific date and time at a specific place.
I guess I could sum up my research in in saying that I found that you really should be demonstrating in your content, the new E by anchoring your at least part of what you’re writing to a specific person or company at a specific time, at a specific time or place, and what you learn from it, and coded learning, so not just like here and at some point I thought, well, maybe just talking in first party will be enough, like I and we. I don’t think it is. I think it’s more about, you know, I we learned x, y, z through ABC process when we were blah, blah, blah in the past, because we’ve been doing this so long.
So now we’re trying to encourage our clients to make sure they include that level of, I’m calling it, Vitamin E in their content. And yeah, it’s not easy. It is not easy. So it’s been, it’s been hard.
David Bain
Great advice. At some point I’d like to move on to talking about tracking, measuring the impact of, obviously, what you’re advising, and managing to get the citations, get the mentions of whatever you’re trying to optimize on AI mode there as well. But, but let’s maybe do that a little bit towards the end.
Shall we talk a little bit more about types of content? Because, Rana, you suggested quite a few different types of content, and maybe if you could perhaps talk about different forms of content that tend to be more effective for optimizing for AI mode, please.
Rana Abu Quba Chamsi
With Gemini and what’s going on now, we are not only searching through words, we are searching through images and through short videos. Search is changing.
My advice, as I have said, I work a lot with small businesses, is to take care of the details. Adding the alt-text to the image is much more important than at any time before, and having a real brand, having your real photos on your websites and being authentic. It’s something that has been basic when the award of SEO for generations, but now it’s more critical, because when people are searching, they are taking a photo, they are looking for your product, your product should have a brand, be unique and and even the font which is related to your brand. This can help to build the trust and all the signals that Gemini will use also to recommend and to find these results for user who is looking for this.
So use context, contextual labeling and the physical environment, your photo, everything, take care of details
David Bain
You also talk about long context optimization for niche authority. What is that?
Rana Abu Quba Chamsi
For each niche, use your intelligence in this domain. What does this mean? It means that for each niche, think what your users need and try to offer the services at the end. For example, instead of putting only contact us, we should provide the API’s that can let us have the appointment directly. Or say I’m looking for recipe for $30, instead of only showing a list of things and the recipe, add the functionality to add them directly to the basket, and I can order this stuff to my house, even for decoration, even for travels.
Always think of your niche and try to offer one more service on your website so that Gemini will prefer this, because it’s much more adequate to the needs of the user.
David Bain
Are users really interacting on LLMs now? And what are you doing right now to improve visibility into LLMs?
Araminta Robertson
So I’m very proud of our framework that we have developed to help clients grow visibility on LLMs. So we call it “GPT Articles”, and it’s quite straightforward. So essentially, the method is that we first decide on what we call bottom-of-the-funnel topics. So these are the money keywords, the money topics, where you know this person searching this up is clearly close to converting.
So if we take Majestic, you know, someone who’s looking up for an SEO tool, that kind of stuff, so we develop these bottom-of-the-funnel topics, and then for each topic, we put together what we call “GPT articles”, which are short articles about 1,000 words, answering very specific prompts. Maybe, like top “SEO tools for XYZ” or “Majestic versus a competitor”, and then developing these short articles that answer the question very succinctly.
There’s a few best practices for LLMs, like, for example, you want a relatively long URL, you want it to be quite factual, have a conversational tone, and to answer the question immediately. Those are the kind of things where we have seen a substantial increase in LLM visibility for our clients.
To give you an example, we had one of our clients who are an embedded payments company, and we took them from like 5% visibility to over 34% and what’s even more interesting, is that they also saw an increase in inbound conversions. So it was really exciting.
I think the hard part of LLM visibility is tracking, because we all use ChatGPT or Gemini and how we’re searching is so fragmented. Now, you might do your initial bit of research in Gemini, but then you’ll go to Google, and then you’ll go to LinkedIn. So it’s really hard to track and attribute those conversions.
With our client as an example, the way we were able to attribute this is in two ways. First of all, we use an AI tracking tool called Peec. There’s many out there, like Profound and all that, and they all do the same thing, so I think they’re all fine, and that allowed us to see the visibility of the client month over month. And then the other side of it was, whenever someone submits a form on the website, we simply asked them, “how did you hear about us?” and those two metrics, they’re not exact and they’re not perfect, meant that we were able to see clearly an increase in inbounds coming from LLMs. That’s kind of what we have found works well.
I also think that alongside Digital PR and increasing links in third-party websites is a really solid strategy for improving visibility in LLMs.
David Bain
Nikki, Araminta talked a little bit about conversational search there as well. So how is that impacting the way that you actually choose to select what kind of content that you want to optimize? But also, how do you optimize for follow up prompts?
Nikki Halliwell
Yeah, it’s changing all the time. We’ve got our own framework that we created called VISIBLE, which you can check out in SEO in 2026, the book that Majestic offers, but, yeah, it’s changed massively in terms of follow up prompts. It’s not just looking at queries anymore, like, what is EEAT? It’s like, how does EEAT apply for my specific industry?
It’s almost like, when you’re creating content similar to what Rana was saying, and being able to offer the additional thing, whatever that thing is that is relevant to that website in terms of the content, you’re trying to take it a step further and anticipate what the follow up question is going to be in the LLM or in in AI mode, because they’re always trying to keep you going with those conversations and do the next thing. So it might be like, what is EEAT? But then, how does EEAT apply within the healthcare niche or things like that?
David Bain
So you anticipate what the follow up question is going to be. Are there any software tools to actually use to do that, or is it just instinct?
Nikki Halliwell
It’s a bit of both, but don’t just try and optimize for the sake of it. For example, if I’ve got a healthcare website, I’m not going to try and target something like how does EEAT apply to law, just make it relevant to you. I have definitely seen instances of people where they’re trying to target everything, which is the typical sort of content that you see certain people posting on LinkedIn, but it’s just using the data that you do have, or lack of data, in this case, and just actually asking your audience “How did you find us?” and having that functionality on the website, and that’s one of the ways that we can get around the tracking nightmare that we have at the minute.
We CAN look at brand mentions for tracking as well, which can be one way to do it, but just trying to look at a lot of third-party signals, like backlinks, and the EEAT and everything else that people have already covered. But how can we influence that so that when we’re having these conversations with AI mode, and it’s ready to check out, because you can check out within AI mode now, so you don’t even have to visit the website. How will we get to the point where it’s my website that’s showing up in there and not my competitors?
David Bain
Charlie, there’s obviously going to be more interaction directly on LLMs now with follow-up questions. So how does AI mode affect the buyer funnel, and what do we need to do when we’re bearing that in mind?
Charlie Marchant
I love this question. So for me, what’s happening with the buyer funnel is something that used to be really manual. When we think about 10 blue links, we used to have to go through and trawl so much information by opening singular web pages when we were right at the top of the funnel, even if we were just asking quite simple questions at that time.
So for example, if I had a problem, I feel like I’ve got dry skin, and I’m using moisturizer, it’s not working. And you Google something like, what are the causes of dry skin? You end up reading multiple articles to actually get there. Now, AI mode and other LLMs like ChatGPT are compressing that buyer a funnel significantly, because they’re able to essentially give you what is almost like a blog post answer that answers that really quickly, without you having to trawl through multiple different links, which is fantastic because it means for businesses by the time someone is actually at the point of hearing about their business, assuming that they do show in AI mode, they’ve already been educated by AI search.
Now that’s not what a lot of the SEO community want to hear. Because for years we focused on clicks and traffic. There’ve been metrics that we’ve used to measure. But when we’re thinking about what actually impacts the business, which is the conversions, the leads, the sales that they ultimately make for the revenue that they are trying to drive.
What we care more about is someone’s buyer journey and making sure that someone gets to them at the right point in that buyer journey. So if we’ve got those conversations going on in AI mode, someone can quite swiftly move down a buyer journey funnel just from putting in an informational query, having a conversation back and forth, ending up, in that example, with product recommendations. You might end up with recommendations and multiple different skincare brands for products that you could buy to support your specific issue with dry skin.
The bit that’s going to become even more difficult here, though, is the personalization. We know that AI mode is personalizing already. It will become increasingly more personalized. Google just very recently launched in the US personalization so that Gemini is able to look at all of your different apps, look at your Gmail, maps, different searches that you’ve done in the past, past product purchases that have come to your Gmail confirmation, all sorts of different parts of the Google ecosystem.
That’s a little bit scary, but also a little bit fantastic at the same time, because then it means the kinds of answers that I’m going to get in AI mode are going to be personalized to me as well. So if it knows what you’ve bought before, then the AI mode is going to be more likely to recommend you products from brands that you’ve already purchased from that you know well, which means the buyer intent can be much higher, leading to a compressed funnel, and hopefully better buyer intent.
I think there’s still some evolution to go from what I’ve seen with my own conversations with AI mode, but until we’re there, the direction of travel, I think, is very, very clear.
Nikki Halliwell
Yeah, I would 100% agree with that. I’d also just add on to it and saying, I think it is completely changing what we would know as that buyer’s journey, isn’t it? And it’s almost getting rid of the consideration stage, as Charlie described, like you’re chatting to AI mode, and they now have the ability to check the specifications, check stock, compare with other brands, and even offer discount codes, all while the user is describing the problem, and then it can even check out for you.
So it’s completely changing the way and will hopefully continue. Well, I say hopefully. Hopefully is debatable, but it will probably continue to change the way us and the customers of our clients check out, continue to go ahead.
David Bain
I was just gonna say Rana, obviously, Nikki and Charlie were talking about personalization, and you shared beforehand, semantic personalization via Google workspace. So I was just gonna ask you to talk a little bit about that, please.
Rana Abu Quba Chamsi
I think Charlie has just talked about that when she said that. Now Gemini has included this capacity to search and all our other stuff, like our Gmail, like our documents, like our calendar, even, and it can see the files that we have in our Gmail. So imagine that we are looking for a program like Asana, and I already have something on my hard disk, so it will recommend this much more for me.
My advice here for brands is do something that people will like to download, for example, like a useful PDF, useful checklist, stuff that you will be found in people’s Gmail’s and in their Google workspace. This will increase your possibility to be recommended when they are looking for something and comparing some products.
One another thing that Charlie has just mentioned, they are comparing information and everything. So be clear in your writing, because it might be you might have the risk that it will be misunderstood, or these LLMs will translate it not exactly in the way that you wanted to show it. So our writing should be clear. Everything should be clear, so that LLMs can understand it as it is, and compare it to other products as it is.
I will give the example of Salesforce and their pricing. They have a very clear table. You cannot misunderstand this as an LLM and give wrong information.
David Bain
Pam, Rana used the phrase “increase the likelihood to be recommended”, which is something you mentioned beforehand, as well as content include for AI optimization, a site trustworthiness checklist, I would imagine that does this as well. What is a site trustworthiness checklist and what impact does it have?
Pam Aungst Cronin
So when I sent you my brain salad of notes that was part of an example of a fan out query. So what I was referring to, is I used an example of “What is EEAT” and in a regular Google search, there’s a bit of a fan out type of follow up prompt sort of thing in the People Also Ask, and People Also Search. And as SEOs, for many years, we’ve been optimizing for that. We’ve been putting those as secondary key phrases in our content subsections of our content.
But if you ask in AI mode, “what is EEAT”, it will A, go much more in depth, and B, the follow up prompts are just so much more in depth. So to compare, I Googled “the importance of EEAT” and the people also ask sections were EEAT examples, so “What is EEAT”, “EEAT checklist”, “how to improve EEAT guidelines”. Just very, very simple and somewhat repetitive.
When I asked AI mode, when I searched in there, “importance of EEAT”, the follow up prompt was, “would you like to see a checklist specifically tailored to improving the trustworthiness of a business website”.
You can see the difference there, but this point tacks on nicely to what was just being talked about, which is to help the user complete their information journey. Be the one that has that checklist, and not just the basic checklist, like in the people also asking EEAT checklist, like in-depth trustworthiness for a business website checklist. And not only will you have a chance of showing up in that fan out query in AI mode, you’ll also have a chance to, and this is the second thing it tacks on to, to land your downloadable assets in the in their Google Drive or their email or whatever, and get your brand on their own little personal LLM worlds radar, which I think will intensify even more over the coming year, because the AI browser wars are here.
If we’re being honest, Chrome and Gemini are going to win the AI browser wars, because they already have the market share. They’ve already put Gemini in the browser. They just need to tightly link these things, just like they are with AI overviews and AI mode.
Another example I have is the nutritional value of salmon. When I search that on Google, you get the typical People Also Ask, “what is the nutritional value of salmon?”, “Is it okay to eat salmon every day?”, “Why is salmon a super food?”. But in AI mode, the follow up prompt was, “Are you looking for specific cooking methods to preserve these nutrients, or would you like to compare the nutritional value of different salmon species, like Sockeye versus Atlantic”. So immediately, if I was on a website that sells and ships frozen seafood or something, I’m like, Oh, wow, I need to make that comparison chart and an article on cooking methods to preserve salmon nutrients.
Araminta Robertson
I think this is a really great way to find content ideas. And I would say that you can then find short titles of articles, answer them, and create content that answers all these sub-questions. I think that is the key to improving your visibility. It’s just a lot of more fragmentation or atomization of basically how people are asking questions online.
You also need to create content in a way that an LLM can crawl. For example, images it can’t crawl. So you want tables and written content texts and to make sure you’re going after the right prompts. Like those ones are great examples, and I think it’s really a good way to find ideas.
David Bain
I was just gonna ask Pam and perhaps anyone else, how do you conduct AI-search keyword research on mass nowadays, with actually having to conduct queries yourself? Are there tools out there that are harvesting new, more relevant long-tail keyword opportunities for AI search?
Pam Aungst Cronin
Not that I’ve seen, but the approach that I’ve been taking with everything lately, is now that AI has made application development so much easier, I’m just going to build a tool for it, like many others we’ve already built for in-house use, at least for now.
David Bain
I’d like to talk about tracking, measuring success just before we do that. Rana, you also mentioned AI native content formats, like NotebookLLM, with a view to actually creating content to optimize?
Rana Abu Quba Chamsi
That has been a little bit covered in the discussion, but not using the same words, it’s about also the same thing. It’s about creating the downloadable content that we have been talking about. It’s about really creating stuff that we will download and put in our Google workspace so that add in our LLM, so that it will make the search more oriented to what we are interested in. And all of this depends on really having high quality content that’s created, and will invite people to really need to install that on their PCs. So that was the idea.
David Bain
Thank you. Nikki, shall we start with your opinion on tracking? I mean, tracking is obviously more of a nightmare with people being on multiple platforms. Is it possible to track and what do you actually do to try and measure success?
Nikki Halliwell
It is possible, but it’s a question of how much you trust it. I mean, we do see some examples of traffic coming from ChatGPT and those sorts of sources within Google Analytics. But then when some of my colleagues dig into it, we’re also seeing some of that traffic that’s actually getting attributed as direct or even unknown, and it just all gets very, very messy.
It’s why one of the ways that we’re trying to get around it is similar to what I mentioned earlier, is recommending that our clients use surveys at checkouts. How did you hear about us? And that gives a good mention and brand tracking I’ve mentioned already as well, but I think, like with Cookie Consent Mode and everything else just makes it all the more complicated and throws up more and more barriers. I think it’d be interesting to see. I can see a lot of people nodding, so it’d be good to see if they’ve had similar problems, which it sounds like they have.
David Bain
I like checking out the “nodders” as well. I think Araminta was the first one to do the nodding there. Why is tracking so hard in and what can be done to improve it?
Araminta Robertson
It’s really hard because our clients hire us specifically for conversions, so we help companies turn content marketing into a channel through SEO and LLMs and so like any other external provider, you kind of have to prove your that you’re doing that and you’re bringing in customers, and that’s where attribution comes in.
You need attribution, because you have to be able to prove that you’re actually bringing in new web form submissions, new leads, MQLs and customers. So that’s why we care so much about tracking. And the way I think about it is it’s super hard. Doesn’t mean that we shouldn’t try, though. I really think it’s worth trying as much as we can and not just give up.
The thing that I’ve seen people maybe get the wrong way is that tracking traffic from LLMs, I don’t think is the best metric, because think about how you use ChatGPT and these LLMs, they’re, to me, more of a discovery channel. How often do you really click on the citation link and click through and head to the website? Not very often, right? And so what tends to happen is you do your research.
Imagine you’re looking for a backlink checking tool and Majestic comes up, you would then go to Google and type in Majestic, and then that’s seen as a branded search, or if you’re doing paid or like direct or whatever, organic. So that’s why traffic, which we do track, but it’s not the most interesting metric.
We have clients where I think 2% of their traffic is coming from AI, but 30% of their inbound leads come from LLMs. That is the difference. So that’s why our two favorite things to track are, as I said earlier, the visibility via a tool like Peec, which shows you your visibility on specific prompts that you track over time. And then the “how did you hear about us?” Which is not perfect, and obviously people get it wrong. But again, attribution is not perfect, so we just work with what we have, and usually those two things allow us to track a little bit the influence of LLMs.
David Bain
Thanks Araminta and a fellow nodder was Charlie. Charlie, what are your thoughts on tracking?
Charlie Marchant
We track in a really similar way to this. So we do track traffic and conversions from LLMs for our clients, you can do that in Google Analytics Four if you set up custom channel groups, we do it through Looker Studio, so that we can filter it into a pie chart and also a graph that shows over time.
That’s particularly useful for us, because when we look at the traffic and conversion trends on that graph, for the majority of clients, they’re increasing over time, they’re working towards improving AI searches, there are some click throughs for certain types of products and services in particular, directly, as kind of Nikki mentioned, there’s been a huge rise, though, in dark traffic, in traffic that’s really hard to attribute, that’s often coming through direct and one of the correlating factors, when you Look at that graph of the AI search traffic, is another increase, which is through direct traffic instead. So you know that it’s eating a share of your pie that is therefore not easy to attribute, but you can see some sort of correlation in most analytics.
That’s the kind of pattern that we see for our clients. And then also, yes, also just having people self attribute where they’ve come from as well, we’ve seen increases in people saying they’re coming from ChatGPT or other LLMs.
David Bain
Rana, you talk about the verified source premium in SGE (Search Generative Experience).
Rana Abu Quba Chamsi
It’s about really increasing trust in our brands. Traditionally, we do this by having great backlinks and great internal links on our website. As Pam has discussed before, most importantly, it’s about putting our real experience in the subject. That’s how we can put our brand so that, if we are talking about financial stuff, for example, maybe an article that has been written by someone who’s saying, I was in a debt situation, and I get out of that through this and this and this, such an article, can have a huge increase, huge ranking in LLMs, and it will be shown more than something has been written by experts, because experts will be similar one to the other one. The idea here is to increase trust through writing, real experience, a real situation that we treat here and there.
For tracking, let me share an answer also to this question, because it’s also very interesting. I feel that the most important thing here is to explain to our clients that the world has been changed. They will be looking for clicks, they will be looking for the old KPIs, and we have to tell them that that has been changed and that we cannot measure like this anymore. So this is the more tricky stuff for us, or at least for me.
Now, I don’t know if you share with me others at the panel, and then we have these tools, like other people trying to measure through the LLMs, but one thing that I find very important, the number of clicks has dropped, but the quality of clicks has increased, which means that when people now are clicking, they have seen us, and they are looking for our brand, which will increase, they are much advanced in the funnel journey. So that’s also a positive point, if we can look at it from this point of view.
David Bain
Pam, would you like to add some final thoughts on tracking?
Pam Aungst Cronin
I would like to agree 100% with what Rana just said, that is like the quality of the clicks like we’re we’ve always reported on organic search conversion rate, and as soon as we added LLM traffic to our Looker Studio dashboards.
Of course, we look at conversions from those two, but it’s more about redirecting their attention because the client does tend to lock on to the traffic metric. We keep pointing out that, although we’ve got this alligator mouth going on where impressions are going up, but clicks are going down, I actually have taken that chart from Search Console and overlaid it with their conversion rate from GA4, and there’s an inverse correlation like that. Yes, you’re getting less clicks, but look at how much better they’re converting so you’re getting more quality traffic, as opposed to a higher quantity of traffic. And that’s definitely been making some of the clients feel more optimistic about all of this.
Another metric that we’ve developed in this absence of KPIs that we have related to AI powered search is AI Overview Visibility percentage, so we track how many total keywords the client ranks for and how many of those have an AI overview available? Because not all of them are triggering them yet. It’s getting bigger and bigger, but not all yet. So how many of those rankings, existing rankings do show an AI overview. And how many of those are you in? And that kind of like, market share, per se, of the AI overviews has been a metric that they a KPI. It shows awesome low hanging fruit opportunities, because if you get a yes, like, we’re ranking for this keyword in traditional search, and yes AI overview exists, but no, that site’s not in it yet. A few small tweaks to that content, and it probably will jump up there very quickly. So our clients are really liking that one too, right?
David Bain
I guess one frustration that SEOs have is the fact that if the SEO based upon the knowledge that they have now and what’s happening in Google AI search now, it’s perhaps going to be several weeks or months until their content is optimized enough to appear, and perhaps there are going to be other elements that occur in the future that they haven’t thought of today.
So where is Google AI search going? What’s it likely to look like in six months time? Shall we finish off with that thought and just reminding the listener where people can find you.
So Nikki, where’s Google AI search going? And what do SEO SEOs need to do to actually optimize for the future today and where can people find you?
Nikki Halliwell
It’s going to become more and more of the default. I mean, a few of the amazing ladies on the podcast today have already spoken about how it’s already integrated into the browser in Chrome, and if you’re using it on mobile, when you open up the Google app in future, I imagine that will be the default experience, rather than doing the traditional search first, it will be AI mode.
Personally, I think there’s just going to be more and more of the personalization and the Agentic Search, and it’ll be interesting to see how that develops with large e-commerce stores and marketplaces, they’ll be trying to do their own version of this. So instead of having to go to AI mode, I’ll go to “Insert brand name here”, and I’ll be looking for a new pair of shoes, and I’ll type in my shoe size and what I want them for. And instead of going to a third party into all the different avenues like AI mode was, it’s keeping me on this one website, and I think there’s going to be more and more of that, but for the time being, it’s much more of that personalization element within AI mode itself.
David Bain
And where can people find you Nikki?
Nikki Halliwell
You can find me on LinkedIn. Just search for my name, and you’ll find me there. My website is nikkihalliwell.com, and my newsletter is techseotips.co.uk.
David Bain
Thank you so much for joining us, Nikki. Charlie, where is Google AI mode going? And where can people find you?
Charlie Marchant
My perspective on this is that AI mode is pretty much a test bed for how Google wants search to look in the future. So I think AI mode, or something similar to it, an AI search experience will become the default in Google over time. But I think there’s a lot of kinks that they’re going to want to iron out first, firstly, making sure that users trust and will use AI mode, which is kind of the system that we’re in now. They’re trying to move people from AI overviews into AI mode, so that users actually start to enjoy the experience of using it.
But also, the other big area that Google is grappling with is how they integrate ads effectively into AI mode as well. So they’ve already launched test ads in beta version in the US. They haven’t rolled out any further than that yet, but 90% of Google’s revenue is from Ads. More than 90% maybe. So there is absolutely no way they will be moving AI mode into default until they are sure that those ads convert really well and will be driving at least the same, if not more, revenue for them.
Our direction of travel, I think, is clear though, and that is that AI search experience is going to become more and more normal, and the days of the 10 blue links are moving much further behind us quite quickly.
Those who want to find me, you can find me on LinkedIn. I’m always happy to have a chat on everything future of SEO, and I also do a weekly podcast called The Dojo, which you can find on the Exposure Ninja YouTube channel or Spotify Apple podcasts and everywhere else that you may get your podcasts.
David Bain
Thanks so much, Charlie. Rana, what’s the future of Google AI search, and where can people find you?
Rana Abu Quba Chamsi
First of all, I have always thought that Google will dominate this world of AI, because Google is Google, and we were just waiting for it to recap, because we have seen OpenAI were the number one for for a while, but now we are so happy with what’s going on in this world of Gemini and what’s going on relating with the AI mode.
In my point of view, I think we will be moving more for the execution AI, which means, I believe that the sites which will work better, it’s the website that will understand their clients and will take them to the end of the journey, providing APIs and letting this Gemini and this Chrome to take their websites from a normal website to execution. So as I have said, we can reserve a restaurant directly, we can take an appointment with a doctor directly, and it will make these tasks much easier for us.
I think this will be number one and number two. It will also increase the search through images and videos, because I believe they believe in that also. So that’s how I see things.
I’m very happy to see you on LinkedIn. I’m active there, and I will be very glad to have a virtual coffee.
David Bain
Araminta, what’s your thoughts on where Google AI search is going over?
Araminta Robertson
I very much echo what everyone else has said. I really do believe that the plan here is for Google AI mode to be the default, right? And it’s the first thing you see when you open up Google, or even in your tab when you’re typing there. And so I think that what brands can do now is you need to start really thinking about how to appear in LLMs.
I think the number one thing is to start focusing on the bottom of the funnel, as people have said here before, when someone types in something like, “what is SEO right?” Google or LLM is just going to answer their question. There’s no brand recommendation. So does it make sense to create content to answer that? Not so much. What you really want to make sure is that you turn up when someone types in something like SEO tools right at the bottom of the funnel, because you know whether you like it or not. If you don’t appear there, they might not even know you exist.
So creating content targeting the bottom of the funnel, I think that’s super important if you want to start getting traffic from LLMs and actually getting conversions. And if you start now, then once AI mode is fully launched, you should be in good hands, and you should be doing well. So I think that’s the number one thing.
If anyone wants to find me, I am primarily on LinkedIn at Araminta Robertson, and you can find us on our website, and we have a newsletter, podcast, YouTube channel, all of that as well, at mintcopywritingstudios.com.
David Bain
Thank you so much. And Pam, where is Google AI search going? And where can people find you?
Pam Aungst Cronin
I echo all of that. I think we’re going to move from AI overviews to AI mode, but then I don’t want people to stop their thinking there. Then we’re going to move to AI integrated browser and then to a Gen Agent, AI browser. So we’re going to get to a point in the relatively near future where things are happening the way that these examples have been described.
The agentic browser is going and doing the shopping and adding to cart and checking out, maybe with just a quick confirmation from the human, but the humans aren’t going to go to the websites anymore. So if you want to think really far ahead, which I think we all should, because this goes so fast, I would think towards agentic user experience optimization. How does your website perform for when an AI agent comes to it?
This is going to be a huge technical challenge, because a lot of websites are programmed to block bots. Anything that tries to do anything that’s not a human on a site is typically at least attempted to be blocked. So testing some of these early AI browsers, I’ve not had any success with their actual utility, because the websites won’t allow the AI agent to do anything. But this is going to become such a demand, you’re going to have to figure it out.
You’re going to have to figure out how to allow an AI agent to fill out a form on your website or to check out a product on your website. The humans just aren’t going to do it anymore. At some point they’re going to get spoiled. They’re not going to do it. It’s going to be like, an analogy I use is like, if you want to make a doctor’s appointment, you can just call, well, maybe if you’re lucky, you can call your doctor on their cell phone and say, I want to see you next week. But most of the time, you have to call the office, and you have to go through the staff. Think about your clients, your target audience, like that. You’re going to have to go through their AI agent to get to them at some point. So the earlier you start thinking that way, the better.
You can find me anywhere online at pamannmarketing.com and all my socials are at the bottom of that, or over on LinkedIn.
David Bain
You’ve been listening to the Majestic SEO panel. If you want to join us live next time, sign up at majestic.com/webinars and of course, check out SEOin2026.com, too.
Thank you so much to our panel today. Hopefully you can join us again soon. Bye for now.
Previous Episodes
Follow our Twitter/X (@Majestic) and Bluesky (@majestic.com) accounts to hear about more upcoming episodes!
Or if you want to catch up with all of our previous episodes, check out the full list of our Majestic SEO Podcast episodes.
- How to Optimize for Google AI Mode - January 23, 2026
- Advanced Filters available across Site Explorer - January 19, 2026
- How to Set an SEO Strategy for 2026 - December 17, 2025






