JD Sports on AI search and digital transformation: we can add ‘so much scale to personalisation’

AI-powered, or vector-based, search is one of the technologies being added to ecommerce stacks to deliver better customer experiences (see our article, ‘What exactly is AI-powered commerce search?’).

Many online retailers are in the process of switching from search functionality that was bundled with a commerce platform over to standalone AI solutions, often as part of a wider shift towards so-called ‘MACH’ architecture (micro-services, API-first, cloud-native and headless).

JD Sports is one growing brand undergoing this move to composable architecture, including AI-search, in what Arianne Parisi, EVP Chief Digital Officer, describes as “a customer-centric transformation” of its tech stack to something that “is a lot more agile” and can “serve consumers globally”. This transformation is intended to power continued growth in the US, home to 225 of JD’s 3,400 stores and where the company has ambitious expansion plans for “80 to 100 new stores per year”, as well as in Europe and APAC.

The brand, based in Bury in the UK, recently announced it would be implementing Algolia’s API-first search solution to improve the browsing and shopping experience on the JD websites via vector-based search. Essentially, the tech provides more accurate and relevant search results, as well as merchandising and recommendations, and opens the door to more immersive or conversational search experiences. Users can ask questions, almost as if talking to a personal shopper with knowledge of the full product catalogue, rather than searching for words and phrases.

Search experience as the “core commerce experience”

The search solution will initially draw on data including JD’s product catalogue, browsing habits and clicking and conversion events.

Parisi describes the experience as one that “adapts and adjusts to be hyper personalised based on user intent.”

“What we will come to market with is something that really maps to the user journey depending on what they’re looking for as an inspiration; are they really intentional about what they want to find within that session, right?,” she adds.

The search experience is “the core commerce experience”, says Parisi, but adds that the connectivity of search, merch and recommendations “is really what creates the magic formula, because we want every touch point to be highly personalised.”

“It should be quite an immersive experience as the customer navigates any of our digital touch points.”

The potential of RAG frameworks for new user experiences

Algolia CEO Bernadette Nixon outlines the potential of large language models in search to create new ways for consumers to interact with a product catalogue and cites shopping guides as an example of a generative AI experience the tech company has created in beta.

“We did some user research of consumers and merchandisers to understand their priorities and what we should develop. And the very first one was shopping guides.”

“So, it might be for an upcoming Mother’s Day or a Father’s Day, for example.”
Nixon says there’s “a hunger from the consumer for that type of information to really guide them.”

“We have JD’s full catalogue, we apply a RAG (retrieval-augmented generation) framework to it as well, so that you avoid the issue of hallucination and then you can serve it up in a different user experience. Or you could make sure your fifth result in any result set is always a shopping guide, for example, and then you can do product comparisons and have a full-blown conversational commerce experience,” Nixon explains.

Though this is just an example and won’t be part of initial rollout, it shows the potential. As Nixon says, “It’s very time consuming for the marketing teams and the merchandising teams within retailers to produce that material. So it’s a perfect example and a perfect application for generative AI.”

The omnichannel vision for loyalty

JD’s implementation of Algolia won’t initially draw on data from its loyalty programme, though Parisi talks about a “future vision… where we can save your preferences and understand the ways you browse and the ways you buy, all within consent of the user, to say, ‘can we make this a better experience for you?’”

“Loyalty is one mechanism for us to get closer to the customer and understand their purchasing habits, online and offline. And ultimately, we want to stitch together that journey a lot more tightly, so that we deliver something to them that’s relevant depending on where they’re browsing and transacting,” adds Parisi.

These relevant experiences can be “at the front end, around research and consideration or even post purchase,” says Parisi. “Ultimately, the vision for JD is that we understand a customer’s behavior. And then the way we communicate with them over time, becomes more and more relevant and personalized. So, what you purchase in store, could affect what you find on the site the next time you land, right?”

“We know that customers, when they when they share their data with us through loyalty, they have an expectation that things become more and more relevant to them.”

“There’s no debate” about the value of personalisation

When asked about the value of personalization, given years of debate in the industry about the relative merits of the idea of 1:1 messaging, Parisi says, “Personalisation is one of the more vague terms, right? And it can be applied in such a diverse number of ways. To me, there’s no debate or challenge to the value of personalization; really it’s just representative of the strength of your data and your ability to action that data.”

“I think we’re going to become a lot more untethered from the ways we used to constrain and define personalisation. AI is where we’re going to add so much scale. And I think in marketing… it’ll be really disruptive. Especially because gen AI can start to innovate with you on content creation.”

As for what this influx of new AI-powered martech means for marketing teams, Parisi says, “It really forces you to better define, upstream, your brand voice. Because you have to then have an expectation that downstream, as that marketing is deployed, it will become prolific. So, it’s really about having personas and standards up front, and around how you input those into the business and into all of your touch points, [so] that it stays relevant and stays on brand.”