HOW IS AI INTERACTING WITH YOUR STORE? MY THOUGHTS ON SEO / GEO / AEO FOR ECOMMERCE
- Dan White
- Jul 21
- 9 min read
Updated: Jul 29
Head's up. This is a evolving brain dump of a post which brings together something of a library for my mind about the current state of play when it comes to AI's and SEO / GEO / AEO - your acronym of choice. You'll also notice that I write AI’s – plural. Not AI singular. Using the singular term assumes all AI’s are the same when they really, really aren’t. For the purposes of this article AI’s covers ChatGPT, Google AI Overviews, Google AI Mode, Claude and Gemini.
Current AI Search Questions
Is information and product management becoming more important?
How do we know what searches we appear for in AI’s outside of branded terms?
How many people are searching for terms we might want to show up for?
Like keyword tracking, can we track our performance in LLM’s?
Can we track what sales and traffic AI’s send to our website?
Who should we follow to keep on top of all things AI in Search?
Do we need to alter our SEO strategy and / or marketing budget?
Note: AI search changes are so fast moving this post will likely be out of date in a couple of months. I'll try and keep it fresh where I can. If there's anything I've got majorly wrong, or you have thoughts, then send me a message.
Do AI’s know about our store?
Yes. Most likely. If you haven’t intentionally blocked AI bots from your website and removed any reference about your brand from the whole of the internet then it’s likely AI’s know something about your brand.
This will come from historical training data, third party citations and using agentic capabilities to still 'read' a website.
What do AI’s know about our store?
All sorts. As AI’s scour the internet and can hold all manner of facts about you. If it doesn’t know a fact it’s highly likely it will hallucinate a fact in response. (Article on why they do that here).
To be sure you will need to interrogate each AI separately asking key questions about the business, the brand and the people within the business to ensure information is:
1) Factually correct
2) Up to date
3) Serves the information with a positive sentiment
I've been building this exercise into my SEO retainer workflows.
Is information and product management becoming more important?
Definitely. I'm planning a post about this. The growing ability for AI's (as well as traditional Google Search) to find and cite ridiculously old and/or obscure information is a newer problem for SEO's. Information cleanup is an increasingly important task.
However, besides risk prevention it has its upsides - this study on PDF optimisation is a good example.
How do we know what searches we appear for in AI's- outside of branded terms?
Most SEO tools are able to pull top level data from the major players (ChatGPT, Ai Overviews etc) but the information they are able to pull is limited. Here’s an example from Patchplants.com from SEO tool SERanking.com showing different topic prompts. Every tool is a little different and data varies wildly.
Bing Webmaster Tools seems to be one of the better sources right now. Google Search Console announced in June that Search Generative AI reports were being introduced although I have not yet had access to them, information that's shared seems to be minimal.
Topic Insights from MS Clarity has also just launched as a way to see how content is performing.
How many people are searching for terms we might want to show up for?
Right now, it’s really hard to say. We get a rough guide from regular SEO data such as Google Keyword Planner and Google Trends data but AI uptake is different according to different countries, languages, demographics and technology use. Sparktoro is a good place to start exploring this.
Like keyword tracking, can we track our performance in LLM's?
We know that people are searching in a much more detailed and much more specific way. Instead of ‘blackout curtains’ people may search ‘give me 5 recommendations for thermal blackout curtains 117cm in length that aren’t made of polyester in a light blue colour’. This is known as the infinite tail.
Tracking the almost infinite number of possibilities is incredibly hard. Sparktoro has also shown that there are wild fluctuations as to what AI’s choose to recommend, so tracking right now could be a fairly ineffective activity: AIs are highly inconsistent when recommending brands or products; marketers should take care when tracking AI visibility
So, how do we get our store featured in a result?
Right now there are 3 areas which may help. Regular SEO, Content Tweaks and a solid Digital Footprint. However, as AI’s are evolving at such a rapid pace a lot of top recommendations are a mixture of hype and guesswork. What we do know is that…
1) Good SEO..
Google AEO is good SEO – John Mueller. Search Relations Team Lead at Google says…
"AI systems rely on search. and there is no such thing as GEO or AEO without doing SEO fundamentals. Tricks will come out and they will work for a short time, companies that want to be around for the long term should focus on something that is proven with long term stability and not tricks."
So a clear, well structured, fast loading website (with a solid XML sitemap) that can be easily crawled and indexed will still be a central component to appearing in AI’s answers. Google even has it in their guidelines: https://developers.google.com/search/docs/appearance/ai-features.
We know also that AI’s use Google’s index for sourcing information: https://searchengineland.com/openai-chatgpt-serpapi-google-search-results-461226 so to appear in ChatGPT it’s likely a brand also needs to be found by Google.
Suganthan has also published this excellent piece on how ChatGPT operates,
2) Evolving Content Structure
There’s a lot of discussion about how AI’s ‘read’ content and correspondingly many websites have adjusted their website accordingly. This includes…
The inclusion of summaries in the first opening stages of a blog post
Including FAQ’s with a clear question and answer structure within posts
Writing content in short declarative sentences
Including more headlines that are more keyword heavy than before
Building focussed pages which seem to outperform long ‘ultimate guides’
There are multiple, often competing studies surrounding this, so it makes sense to create a piece of content which answers everything a user could want to know. One recent excellent study.(which is also a great example of a content structure for AI).
2.1 Adjusting Content Depth
If people are searching in a much more detailed way and AI’s are able to pick out specific references we need to make sure that blog, products and collection pages talk specifically rather than generally. Using the example above if someone is searching for light blue thermal blackout curtains then we need to make sure our product page specifically includes references to all these things to maximise it’s chances of being featured.
(For products especially some of this can also be achieved through structured data).
There is a good study here for how this can be manipulated but shows the power of having items referenced on your website. An awful lot of websites miss this. Why would an AI reference the curtains you sell if you don’t list what fabric they are made from?
Steering what content you need to create and adapt can be done with a range of AI tools including - https://waikay.io/ which will compare AI knowledge bases and compare you to competitors.
Basically, think more widely and more deeply about the content you can use to support product sales. This could be the creation of buying guides, product comparisons, how-to articles, expert advice, troubleshooting guides, checklists and of course, frequently asked questions.
3. Enhanced Digital footprints
As AI’s scour everything on the internet they can to generate answers, digital marketing is becoming a lot broader in order to ensure information across multiple touchpoints is both present and accurate and presents a brand in the best possible light. This could include:
Featured articles on other websites / References in YouTube videos (and YT descriptions) / Features on podcasts / Contributions to news stories and articles / Product write ups and reviews / Listicles / Social media (Instagram was able to be indexed in July 2025. This study shows that AI Overviews cited Facebook 19.5 million times)
Basically brands need to be everywhere that’s relevant so the easiest solution for this is to be working more heavily with PR professionals in order to make this happen. Although the target audience may not be exact it’s thought these mentions should contribute to information an AI holds about a brand.
Anything else?
Is Technical SEO still important then?
100%. It's damn important. From basics like orphaned products and XML sitemaps to accruate internal linking structures, we still want stores to be discoverable as possible. This means that they load fast, have accurate structured data and correct canonicalisation structures (to name a few).
Can't we just use AI to write all our content?
Just because you can doesn't mean you should. In short, it doesn't seem to help...
How often will AI's include a link to us?
It really varies. Different systems are constantly changing when they choose to link out to websites and if so, how they do it. The majority of the time you’re more likely to be mentioned rather than linked to. This can be useful for brand awareness but explains why as traditional Search has evolved, overall traffic levels have reduced.
However, this study from SimilarWeb indicated that brands recommended by ChatGPT were 2.5x more likely to receive a site visit within 7 days, even without a direct click: https://www.similarweb.com/corp/the-downstream-impact-of-ai-visibility/
Isn't SEO and AEO/GEO the same thing?
Does social media influence what answers AI's generate?
Increasingly so. We know that Instagram became accessible to Google's index in in July 2025. Seperately, this study shows that AI Overviews cited Facebook 19.5 million times. And then there's Reddit which as we all know appears pretty much everywhere for every answer. As a result it means that our roles as SEO's need to expand more deeply into what content is shared on brand channels.
Can AI's distinguish between paid and organic features?
When it comes to items like advertorials, it seems that AI's don't really do a great job at spotting the difference. (To be fair Google isn't great at this either). As a result, paying for features to boost publicility to showcase products and influence how they're talked about is more of an important tactic than before. The only difference is that we would expect organic features to stick around and paid features to be taken down once the sponsorship is over - meaning there may be an expiry date on how long that information continues to influence AI's.
Can we track what sales and traffic AI’s send to our website?
Yes, this information is already being collected by GA4 / other analytics packages. We can also see the number of conversions (sales / enquiries) traffic from AI’s can directly deliver. This is a very good post from Kyle Rushton McGregor about setting up further AI tracking in GA4.
If AI's mention a brand and then that translates into a purchase later on then the source would be hidden but could be attributed to Direct or Organic channels.
Who should we follow to keep on top of all things AI in Search?
Do we need to alter our SEO strategy and / or marketing budget?
Quite possibly. Speak to me for a free 30 minute no obligation chat if you have an ecommerce store that may require in search direction.