As AI moves from answering questions to making recommendations and, eventually, purchases, the relationship between brands and customers is set to change. In a Q&A with InternetRetailing’s Managing Editor, Amanda Vlietstra – and ahead of his appearance at CustomerX next week – Ian Macfarlane, acting Global SEO lead at H&M, discusses the rise of the AI-empowered customer, the growing importance of GEO, and why brand trust matters more than ever.
Q: You are one of the speakers at CustomerX’s CommerceAI thinktank “The AI-Empowered Customer” on 14 October. It’s set to be a really interesting discussion – but what exactly is meant by the term “AI-empowered customer”?
Ian: We’re going to be discussing this huge shift that AI agents are bringing to ecommerce in particular. It’s no longer about selling directly to the end consumer. Brands are going to be increasingly disintermediated in their relationships with their customers by AI robots, and that’s an enormously scary prospect for many brands – particularly established brands, which have got more to lose.
It really gets to the heart of the question: What does it actually mean to be a brand in today’s world? What does it mean if your consumer is not necessarily another human being, but an alien artificial intelligence? You’re not just using AI; you’re marketing to an AI. That’s a really interesting question for brands to grapple with.
How will agentic shopping affect how retailers use SEO? Are they now optimising for SEO, GEO and searches directly on websites?
Ian: GEO is clearly here to stay. In the early days, just after ChatGPT came out, everyone seemed to think AI was going to replace SEO because search was dead. Fast forward a couple of years and the picture has completely flipped. AI is here to stay, but we now have a much better understanding of what makes brands appear in AI search – and it turns out, the people with the best ability to do the actual work required to make your brand more visible in AI search are your SEO team.
SEOs have been working cross-functionally with other teams since forever. Those teams might be broader in scope now, with an increased focus on brand PR and brand reputation more broadly, but the core role of SEO is still relevant. We need to understand what needs to be done, how to do it, which teams to engage with, help them understand the value of it and what to do, and then ensure it actually gets done.
Today, we absolutely need to be optimising for traditional search, GEO and on-site search. But I’d argue it’s broader than that. SEOs need to optimise wherever that brand’s consumers are searching. Depending on the sector, that might include social, retailer search or local search. So I’d say those three and more.
What do you expect to see in the short term compared with the longer term as AI search becomes embedded, and how should retailers prepare?
Ian: It’s helpful to see this in the context of the wider arc of how search has been evolving. Initially, we were trying to appear at the top of the classic ten blue links. Then there were different formats to optimise for, like shopping and maps, and then answer-box optimisation. That was really a precursor to AI search. Now we’re looking not only to be found, but to be recommended by search platforms.
At each stage, brands have become increasingly disintermediated from their consumers and have had even less direct control over how their brand is shown and represented. You only have to look at the latest changes Google has made, where AI Overviews appear for searches about a brand and give an AI summary, including pros and cons. That’s a huge reputational thing to grapple with.
And then looking forward, where agents actually purchase on the consumer’s behalf, what does it mean if you’re marketing not to another human being, but to an artificial intelligence that’s making decisions for a consumer? It’s potentially a lot more rational and less emotionally targetable. The things you need to do to market towards an AI are very different.
Something I would like to emphasise, however, is that consumer behaviour takes time. Even I’m not yet comfortable giving what is essentially an overenthusiastic robot intern my credit card and letting it run wild. It’s too early to know for sure how fast consumers are going to adopt agentic commerce in practice. Everyone knows AI gets stuff embarrassingly wrong at times, so whether people are going to trust it with their credit cards right now is an open question.
Retailers definitely need to be exploring this and mapping out what they need to do to integrate with these technologies, so if this does really take off they’re not going to have to wait three-plus years to launch.
Five years is just too far ahead to plan, however. Marketing is changing faster than it’s ever changed, and trying to predict what’s going to happen in AI five years out is a fool’s game.
A lot of people enjoy clothes shopping – I know I do! Do they really want to hand it over to a robot?
Ian: Fashion is very personal. AI, in theory, is a very personalised search environment compared with classic traditional search. AI chatbots are heavily personalised and potentially have access to all sorts of information about you, including what’s in your emails and what you’ve bought.
At the moment, AI is definitely usable as a recommendation engine and as a shortcut for the research process. It will read reviews, look at what people are saying about different products and come back with recommendations. ChatGPT has really pivoted towards that, even though checkout itself maybe turned out to be a bit harder for them to do. Recommendation is absolutely somewhere AI platforms are working well right now.
I needed a new phone earlier this year, and I gave the AI a list of specific requirements. It came back with a recommendation, and that was my shortest shopping journey ever. I would have spent loads of time doing that research myself. There’s a level of trust that it’s going to give you a mostly right answer.
Consumers make that decision based on the problem or the risk of getting it wrong. For a low-stakes purchase, you might not want to read 300 reviews. You might just want someone to say, “This is probably going to be good for you.” If it’s a car, you’re probably going to do something more in-depth. The type of purchase matters.
Brand really matters still, too. It’s a mental shortcut. If I’m given a brand I know and trust, it validates that it should be in my shortlist. I’d be reasonably confident it would be good enough, based on the brand’s reputation and my experience with it. If it was a no-name brand, I’d probably have to do additional research. Brand is still really relevant in the world of AI and shopping. And a brand being well known, liked and respected is also one of the biggest drivers of AI search visibility.
There has always been a level of experientiality to clothes shopping. As a teen, I’d go shopping with my best friends every Saturday. It was a social occasion as much as a shopping trip. But how important are physical stores in an AI-empowered world?
Ian: In an increasingly digital world, the real world perhaps matters more than ever. People are looking for non-digital experiences, whether that’s phone-free nights out or dining out more often. For fashion retail, for some consumers, nothing can replace physically feeling and touching clothes. It depends on the individual consumer, their context and the garment. You’re probably happy to buy a pair of socks online, but if a garment is tricky to fit, you might want to try it on.
For certain brands, physical stores can provide an elevated experience and give people a reason to visit the high street or shopping centre. Done well, that can generate social and press coverage. But that’s somewhat easier for a high-margin brand with stylish flagship stores than for a retailer selling less glamorous products. Regardless, offline will still have its place for many brands.
It’s more important than ever for brands to have their physical-store information correct and up to date online – not just in Google Maps, but in the different sources that Google, ChatGPT and other LLMs draw from. They can only tell customers what they know. Incorrect information can mean disgruntled customers turning up to find a store closed, or not finding you at all and going directly to a competitor. As ever in marketing, get the basic fundamentals in place first.
Could AI lead to a shift from retail categories towards occasions and moments?
Ian: Occasions and moments have always been relevant to retailers, especially in sectors like fashion retail. The key difference is the increased capability AI brings compared with traditional search. AI can answer much more complex questions about specific moments or occasions. In the past, longer searches gave you worse results than shorter ones. That’s completely flipped: the more detail you give it, the better the answer.
It’s also important for marketers to remember that we’re not your average retail consumer. Everybody’s on their own personal journey towards increased AI familiarity, and it takes time for people to understand what they can do with it. Many average B2C consumers’ searches in ChatGPT are still very similar to traditional searches, but the thing that is most often added is the searchers context, and that is indeed often around specific occasions and moments.
Brands should understand which occasions and moments are most relevant to their customers, dig into those consumer needs, and understand where AI platforms are sourcing answers for those needs. Then they can target them through PR outreach, on-site content or other approaches. It varies by sector and by customer. With any good marketing, it really comes down to understanding your customers – and that’s a part of marketing I hope will never change.
Hear more at CustomerX
Ian will be exploring the impact of AI-powered search, agentic commerce and the AI-empowered customer in more detail at CustomerX next week.
To join the conversation and hear from Ian and other industry experts, book your place at CustomerX today.
