UK Retailers Bet Big on AI-Personalised Marketing Ahead of Christmas Trading

John Lewis's £800 million multi-year transformation programme now includes direct integration with Google Gemini and ChatGPT, announced in March 2026 as part of a broader push to make its products discoverable and purchasable through AI shopping assistants rather than just its own website. It's one of the clearer signals yet that UK retail's Christmas trading strategy this year isn't just about better email targeting — it's about whether a retailer's products show up at all when a shopper asks an AI assistant to find them.

Personalisation has moved from the website to the loyalty card

The most visible shift has happened inside loyalty schemes. Tesco has invested specifically in AI partnerships to personalise Clubcard content and promotions down to the individual member, while Sainsbury's Nectar expanded AI-driven personalised pricing across thousands of products through 2025, according to retail trade coverage tracking the sector this year. These aren't cosmetic tweaks to existing programmes — loyalty schemes that once just accumulated points are being rebuilt as AI-first engagement platforms, with the model doing the work of deciding which offer a given shopper sees and when.

The market data backs up the shift in spending priority: the UK AI-in-retail market was valued at roughly $555 million in 2025 and is forecast to grow at an 18% compound annual rate through 2034, with personalisation, demand forecasting and inventory optimisation cited as the biggest drivers of that spend, per market research firm IMARC Group.

The gifting problem, and why it's a harder prediction task

Gift shopping breaks the assumption most recommendation engines are built on — that a customer's past purchases predict what they'll buy next. A shopper buying a gift is, by definition, shopping outside their own usual categories. Retailers investing most heavily in AI personalisation ahead of this Christmas have focused specifically on that gap: building models trained to infer a likely gift recipient from subtler signals, like browsing patterns that don't match the shopper's own purchase history, rather than relying on the same logic used for self-purchases the rest of the year.

Consumers are already ahead of some retailers

A CI&T survey of 2,000 UK and Irish shoppers, reported in Retail Focus in February 2026, found 61% of consumers had used AI while shopping, with 53% doing so often — and flagged a meaningful gap between what shoppers now expect from AI-driven recommendations and what retailers are actually delivering. Separately, a Metapack and Retail Economics report found 80% of UK retailers expect online sales growth in 2026, with AI adoption named as a key driver, and roughly half of UK adults under 45 already using AI tools for tasks like product research and price comparison.

The agentic-commerce wrinkle nobody quite has a playbook for yet

The newest complication is that some of this shopping is starting to happen without a human directly browsing at all. Shopify's Agentic Storefronts feature went live for eligible merchants in March 2026, opening participating stores to ChatGPT's user base directly, and Debenhams Group became the first UK retailer to adopt PayPal's agentic-commerce checkout stack in February 2026, with a wider UK rollout planned for later in the year. For a retailer, that raises a genuinely new question alongside the traditional personalisation one: not just "what does this shopper want to see," but "will an AI agent shopping on their behalf even find us."

The line most retailers are trying not to cross

Retail technology leads have been notably more comfortable talking about AI that improves relevance — better product recommendations, better timed offers — than about AI that varies price between customers, which remains the more contentious and more closely regulator-watched application of the same underlying technology. Getting that line wrong during the highest-stakes trading weeks of the year is a reputational risk most retailers seem keen to avoid, even as they lean harder into personalisation everywhere else.