Sophie McCallum, Senior Producer at RetailJam, caught up with Shaun Loughlin, Digital and Technology Director at leading equestrian brand LeMieux, to talk all things innovation, AI, and customer connection. From virtual try-ons for horses to building a branded social commerce community, Shaun shares how his team is turning bold ideas into reality while staying grounded in what their audience truly wants.
Shaun Loughlin, Digital and Technology Director, LeMieux
Could you start by telling us a bit about yourself and your role at LeMieux?
I’m the Digital and Technology Director at LeMieux. I oversee our entire digital presence, technology, and data infrastructure. That includes eCommerce, B2B marketplaces, CRM, data, and our wider tech stack.
LeMieux is the leading equestrian brand in the UK, covering equestrian equipment and lifestyle products, everything from horse rugs (yes, horses wear rugs!) to children’s cuddly toys in the shape of horses. Hobby horses are one of our biggest growth categories. We’re the market leader in the UK and rapidly becoming a global leader in the industry too. We’ve just received a Royal Warrant from King Charles, so we’re now an official supplier to the House of Windsor.
What’s been the most exciting digital innovation you’ve worked on recently?
It’s not live yet, but we’ve been experimenting with large visual models. My experience with AI in the past has been a bit underwhelming, especially for a business at our stage, but with LVMs (large visual models), we can now do things that simply weren’t feasible a year ago, let alone three years ago. That’s really exciting.
One concept we’re working on is a virtual try-on, where customers can see a product on themselves. The tech exists and works, we’re currently refining it. What’s even more exciting is that no one has done a virtual try-on tool for horses yet. At the moment, it’s all for human body shapes, but we plan to be first in the equestrian space. Yes, you’ll be able to “try it on” your horse.
A lot of companies roll out things like chatbots to save money, whether customers want them or not. In our case, it’s different. Customers are asking for this functionality.”
Virtual try-ons and chatbots used to be clunky. What’s changed to make this more viable now?
Absolutely. I’ve been burned before. I remember trying out chatbots and virtual try-ons 10 to 15 years ago. They were really frustrating, but something has shifted. The technology behind this stuff has moved on. Those old virtual try-ons were based on AR, which, like the metaverse, was briefly the “next big thing” five minutes ago. LVMs do it so much better than AR ever could.
It’s the same with chatbots. Previously, they were basic machine learning algorithms. Now, they’re genuinely AI, neural network-powered. That changes the game.
What about the customer side? Are people ready to trust and adopt these technologies?
That’s a really important point. A lot of companies roll out things like chatbots to save money, whether customers want them or not. In our case, it’s different. Customers are asking for this functionality.
We built an outfit builder using 3D-rendered products: different outfit sets to help people build the perfect matching look. Customers came back and said “this is great, but I want to see it on me and on my horse”— the technology we’re now using lets us deliver that. It’s a totally different situation. They’re pulling for it, not having it forced on them.
How are you approaching social commerce and building community engagement?
It’s similar. We’ve got a huge community around the brand, but we don’t control it. Most of the conversation happens in places like private Facebook groups. Some of those LeMieux-focused groups have 50,000 or 60,000 members, and we have no presence in them.
So, we’re launching an app-based store with a built-in community element, it’s effectively our own social channel. It’ll be integrated with a rewards programme and allow us to foster that community without disrupting it. Think of it like an old-school forum. Brands like Sephora and LEGO have done this really well.
The community can interact with each other, we can gather insights, and (importantly) we can moderate it. We’ll also invite influencers and athletes to spend time there, which will naturally draw people in.
The focus has to be on having a good product and strong reputation. That’s what surfaces you in AI-driven search results.”
How are you preparing for the next wave of discovery, especially with the rise of generative AI?
To be honest, we’d already made many of the necessary adaptations due to Google’s shift towards EEAT (Experience, Expertise, Authoritativeness, Trustworthiness). Generative AI is really just accelerating that change.
It’s no longer possible to “game” the system with SEO tricks. When someone uses AI to search for “the best horse rug”, the model will scour the internet for genuine signals, be that reviews, brand credibility or what people are saying. So now, the focus has to be on having a good product and strong reputation. That’s what surfaces you in AI-driven search results.
Have there been any false starts or learnings in your AI journey so far?
Definitely. When the AI hype first hit and we could build services on top of ChatGPT, we tried using it for translation. We invested quite a bit of time, money and effort into that, but because of our niche and the specific technical terminology in equestrianism, it just didn’t work.
To be fair, we didn’t realise that quickly enough, but we’re now revisiting it. The models have improved, our understanding of how to use them has matured, and we’ll soon be launching several well-localised European storefronts. The heavy lifting on translation has finally become viable.
Do you think AI companies are over-promising on what their tools can actually do?
Yes. OpenAI, Perplexity, Anthropic… they all have a vested interest in making their tech seem more advanced than it really is. It creates a FOMO effect that gets everyone jumping in. Apple recently released a report called The Illusion of Intelligence, which debunks a lot of what these companies claim. These models aren’t reasoning, they’re pattern recognition tools. They’re unbelievably good at that, but let’s not confuse it with actual intelligence.
That’s why I’m cautious about so-called “agentic” AI. I think it’s coming, but it’s further away than people want to admit.
Lastly, how do you make sure you’re building what customers genuinely want and not just what you think they want?
We run a huge open-ended customer insight survey each year. Seven completely open-ended questions with no multiple choice or leading answers. It felt like the right thing to do until we realised we’d received 28,000 written responses!
Six months later, when we’d finally crunched the data, we had a clear view of what customers care about. And because we didn’t prioritise answers based on how they were worded, we focused instead on understanding the real problems behind them. That’s how we landed on the virtual try-on and outfit builder concepts, they answered multiple related needs customers had expressed in different ways. That’s why they rose so clearly to the top.
It’s a lot of work, but we’ll do it every year now and it’s been absolutely worth it.
Discover more…
