TL;DR: AI SEO helps improve your brand’s chances of appearing in AI-generated answers and recommendations. At Hot House Digital, we’ve built AI search into our existing SEO strategy, technical SEO, content and digital PR work, helping clients improve their visibility while keeping track of what’s actually working.
Your website ranks on Google…so why is ChatGPT recommending your competitor?
You’ve spent years climbing Google’s rankings, and things are looking good. Your content’s performing, you’ve got the technical side sorted, and customers are finding you through organic search. But when someone asks ChatGPT to recommend a business like yours, it suggests your competitor instead.
Ouch.
Appearing on page one of Google is still a big win, but it doesn’t guarantee you’ll get a mention when someone turns to AI for recommendations.
For us at Hot House Digital, that’s meant taking a fresh look at SEO. We’re still doing plenty of the things that have worked for years (sorry to disappoint all you LinkedIn gurus), but we’re also looking at what influences AI-generated recommendations and how we can improve our clients’ chances of appearing in them.
And naturally, we’ve been doing a bit of experimenting ourselves. Here’s what we’ve learned and how it’s changing the way we work.
First things first, what is AI SEO?
AI SEO is the process of improving your brand’s visibility in AI-generated answers – whether that’s a recommendation from ChatGPT or a mention in Google’s AI Overviews.
You might have heard terms like generative engine optimisation (GEO) and answer engine optimisation (AEO) being thrown around, too. There are slight differences, but the goal’s much the same: getting your brand noticed when people turn to AI for answers.
The good news is that traditional SEO still does plenty of the heavy lifting. A website that’s easy for search engines to access, with genuinely useful content backed by real expert insight (and we’ll never shut up about the importance of that last bit), gives you a solid starting point for AI search too.
And that’s exactly how we approach our AI SEO solutions at Hot House: building on the work we already do rather than treating AI search as a completely separate job.
How we’ve adapted our SEO strategy for AI search
Keyword research will always be our ride or die, but we’ve added another layer to the process: finding out what potential customers are likely to ask AI tools about our clients’ products or services.
These prompts can get pretty specific, so we’re using specialist software to see which brands are popping up in the answers. If a competitor keeps getting recommended while our client is nowhere to be seen, we want to know why. Is there something they’re explaining better on their website? Or getting coverage in publications that AI tools regularly reference?
It’s a little bit more detective work, but it helps us work out where to focus our efforts. And once we’ve identified the prompts that matter, we need to keep tabs on how our clients are performing against them.
As much as we’d love to spend our days asking ChatGPT hundreds of questions, we’ve got software for that. We’re now tracking brand mentions and website citations alongside our usual Google rankings, keeping an eye on which competitors appear and how the results change over time.
But a ChatGPT mention doesn’t automatically mean someone’s visited your website, let alone become a customer. That’s why measuring your AI marketing efforts means looking at AI-referred traffic, too.
We want to know whether that visibility is bringing people to our clients’ websites, rather than treating every ChatGPT mention as a reason to crack open the champagne.
Technical SEO – making sure AI can actually read your website
You could have the best-optimised website in the world, but if AI crawlers can’t access it, you’re not giving yourself much of a chance of appearing in their answers.
Getting search engines to crawl our clients’ websites is already part of the job, but now we’ve got a few more bots to keep happy. ChatGPT and other AI platforms have their own crawlers, so we’re checking whether they can get to the pages we want them to find – or whether something in the site’s setup is accidentally stopping them from finding the content.
But that doesn’t mean ripping up everything we know about technical SEO. As our SEO Manager, Joel Boden, explains:
“Now that the fearmongering around AI search has died down, and it turns out SEO still isn’t dead, who’d have thought, eh? One thing we’ve learned as an agency is that if you were doing traditional SEO well, such as making sure technical best practices are followed and creating content that is not just backed by search data but is actually useful or interesting to the target audience, then you’re doing the majority of what you need to do for AI search, too. So, rather than needing to make large-scale changes to your process, you just need to adjust and make sure the changes you make are working harder, especially when it comes to content.”
Content – making pages more useful for AI search
Of course, all that research is only worthwhile if we do something with it.
Our SEO content writers have always worked on making content genuinely useful to readers, but AI search has given us another thing to consider: how easily can AI tools find the answers they’re looking for?
We’re using what we’ve learned to take another look at our clients’ existing content. Are their services explained clearly enough? And do their product pages give customers the information they’re actually looking for, in a way AI tools can make sense of?
Take our recent AI search optimisation work for Aevi, a global payments technology company. We researched the prompts potential customers were asking and used those findings to update 13 transactional and informational pages across their website.
One was an older payment gateway article, where we added a short, easy-to-cite summary at the top, clearer explanations and dedicated sections answering the questions we’d identified through prompt research.
And it seems to be paying off! Across 18 tracked prompts with at least five weeks of data following our content updates, Aevi’s AI citation rate increased by 385%. For AI searches related to its payment gateway article, the citation rate climbed from 2.80% to 13.58%.
It’s not just B2B content we’re seeing results with, either. As our SEO Manager, Gabriela Whitehead-Scott, explains:’
“We’ve seen really good results with Product AI optimisation for ecommerce sites, with optimised pages driving 83% of product citations within a 5-month period.”
Digital PR
This is where having a digital PR team under the same roof comes in handy, explained by our Digital PR Manager, Giorgia…
When someone asks ChatGPT to recommend a business, it’s often looking at what other people are saying about that brand. In fact, Muck Rack’s analysis of more than 25 million AI-cited links found that around 84% came from earned sources, compared with just 0.3% from paid content.
That makes the coverage we’re securing for clients even more interesting, especially when it includes expert commentary or original research. And with much of the journalism AI tools cite being less than a year old, keeping that coverage coming is worth thinking about, too.
We’ve seen just how interesting this can get with Aevi. After running a consumer research campaign on trust in biometric payments, we did a bit of digging to see where the research was turning up in AI answers.
In Google’s AI Overviews, Aevi appeared for most of the questions we tested, usually as the first featured source. Its research was even helping shape the opening answers, sometimes appearing ahead of the likes of Visa, the BBC and GOV.UK. Not bad company to be keeping!
ChatGPT was a different story. While it regularly mentioned Aevi in its answers, the sources it linked to were the press release or media coverage we’d secured, rather than Aevi’s own website.
For us, it’s another reason to keep SEO and digital PR closely connected. We want our clients’ websites giving AI tools useful information to work with, while their media coverage gets them talked about elsewhere. We’ve seen the benefits of combining both disciplines in our wider work with Aevi, and we’re now seeing how earned coverage can play a part in AI search, too.
So, where do we go from here?
There’s plenty we still want to figure out about AI search, especially as the platforms keep changing how they work (although that’s nothing new – we’re looking at you, Google).
We’re under no illusions that we’ve got every answer, but we’ve definitely got plenty of questions worth investigating. And with AI search now part of the work we do across SEO strategy, technical SEO, content and digital PR, we’ll keep putting what we learn to good use for our clients.
So if your competitors are getting recommended while you’re nowhere to be seen, it’s probably time to do something about it!
