AI Shopping in UK Retail 2026: Inside the ChatGPT, Perplexity and Gemini Land Grab

Between November 2025 and April 2026, UK retail's front door quietly moved. It is no longer google.co.uk and it is no longer the high street — it is the answer box inside ChatGPT, the chat pane inside Perplexity, and Google's AI Mode tab. Shopify's Agentic Storefronts went live by default on 24 March 2026, PayPal + Perplexity Instant Buy is shipping UK orders via Debenhams, and John Lewis has committed its £800m transformation programme to Gemini and ChatGPT. If your products aren't discoverable there, you are invisible to a growing slice of UK demand that hasn't yet shown up in your analytics.
Discover how AI Mode transforms UK shopping and seize the opportunity to enhance your brand visibility.

The 2026 reset: UK retail is now sold inside chat

Four things collapsed the old model in under six months. The front door of UK commerce has moved from google.co.uk and the high street to the answer box inside ChatGPT, the chat pane inside Perplexity, and the AI Mode tab Google is now defaulting into. The shift is not a forecast — it is a list of launches that already happened.

  1. PayPal + Perplexity Instant Buy launched just before Black Friday 2025, making Perplexity a fully-functional storefront with PayPal as the commerce engine and the retailer as merchant of record.
  2. Shopify's Agentic Storefronts went live for every eligible US Shopify store on 24 March 2026 — ChatGPT, Microsoft Copilot, Google's AI Mode and Gemini, all connected by default from the Shopify admin. AI-attributed orders to Shopify stores are up 11x since January 2025.
  3. Debenhams Group became the first UK retailer to adopt PayPal's agentic commerce stack (February 2026), with a full UK rollout earmarked for later in the year.
  4. John Lewis announced in March 2026 that it would become one of the first UK retailers to fully integrate AI-powered shopping via Google Gemini and ChatGPT, backed by its £800m multi-year transformation programme and an extended Commercetools partnership.

The UK e-commerce market is running at roughly $317bn (≈£250bn) in 2026 and agentic traffic is on a trajectory from sub-1% of commerce sessions in 2025 to a projected 15–25% by the end of 2027. The land grab is happening now, and AI channels are being wired up faster than most marketing teams have scheduled their next review meeting.

Shopify Agentic Storefronts: every UK Shopify store is already in ChatGPT

The most important sentence in UK e-commerce right now: if you are on Shopify and eligible, your catalogue is already discoverable in ChatGPT by default. No app, no integration, no transaction fee beyond standard processing. Shopify gave 5.6 million merchants a distribution channel into ChatGPT's 880 million monthly active users overnight.

Purchases don't complete inside the chat — OpenAI retreated on Instant Checkout — they route back to your storefront via an in-app browser on mobile or a new tab on desktop (see our deep dive on Shopify's ChatGPT rollout). So the agent does the product discovery; you still own the conversion. That makes product data, schema, reviews and inventory feeds the highest-leverage work you can do this quarter. An AI agent can only surface what it can parse.

Perplexity + PayPal + Debenhams: the first UK agentic checkout

Perplexity's shopping experience, powered end-to-end by PayPal for identity, checkout and buyer/seller protection, already lets a US user go from "what coat should I buy for a Manchester winter" to paid order without leaving the chat (see the Perplexity-PayPal mechanics we broke down here). Debenhams Group is the first UK retailer on the rail, with a broader UK rollout scheduled for 2026.

The strategic implication: in an agentic checkout, the retailer stays the merchant of record — so the customer, the order data and the repeat-purchase signal belong to you, not to the LLM. The brands that treat Perplexity and ChatGPT as distribution channels (not threats) are the ones who'll have a CRM full of AI-sourced customers by Christmas.

John Lewis, Tesco and M&S: legacy UK retail goes agentic

  • John Lewis is surfacing product edits in Google Gemini and ChatGPT, has opened a TikTok Shop pilot, and is re-platforming its commerce stack on Commercetools to serve those agent requests at speed.
  • Tesco has rolled an AI shopping assistant in its app to 280,000 colleagues for pre-launch testing, using Adobe's agentic AI stack against Clubcard data to turn 20 years of purchase history into personalised baskets and meal plans.
  • M&S and Boots are both investing in hyper-personalised loyalty reward schemes driven by AI — the loyalty layer becomes the data moat the agents can't replicate.

If you are a smaller UK retailer watching this happen, the lesson is not "spend £800m like John Lewis". It's that the incumbents are already treating agentic commerce as a channel — you need to decide whether you appear in it, or whether the incumbents define it for your category.

OneOff: what a UK AI shopping agent actually looks like

OneOff, a fashion-tech platform we've covered in detail on the Anjin blog, is the clearest worked example of where this is heading. A shopper types a natural-language prompt — "dress like Zendaya at the Venice festival" — and OneOff's AI surfaces shoppable matches from an aggregated inventory of more than 1 million SKUs across Ssense, Net-a-Porter, Mytheresa, Moda Operandi and Revolve, with one checkout.

It is the Spotify-for-fashion model: aggregated inventory, AI-led discovery, unified transaction. The brands that appear in OneOff's results aren't the biggest — they're the ones with the cleanest data, the richest imagery, and the product attributes an AI agent can actually reason over. That is the game for the rest of 2026, whether your category is beauty, grocery, furniture or B2B.

Hidden opportunities in AI Mode (and how to earn them)

Google's AI Mode is not a search layer, it's a shopping layer. It reads Product schema, Review schema, Offer schema, Brand schema and increasingly MerchantReturnPolicy and ShippingDetails. Brands without structured data are simply unavailable for the agent to cite. The same is true of ChatGPT (via Shopify's product feed and OpenAI's Actions/GPTs), Perplexity (via PayPal's merchant directory plus web crawl) and Gemini (via Google's Merchant Center feed).

So the "hidden" opportunity is embarrassingly unhidden: fix the feed. Tighten the schema. Answer the questions the agents are already being asked. The brands still obsessing over blue-link SEO in 2026 are the ones losing the most surface area.

Strategic steps to enhance AI visibility in 2026

  1. Audit which AI surfaces you already appear in. Run your top 20 buying-intent prompts through ChatGPT, Perplexity, Gemini and Google AI Mode. Note where you appear, where a competitor appears, and where a marketplace (Amazon, John Lewis, a Shopify aggregator) is eating your brand query.
  2. Fix the feed. Product, Offer, Review, Brand, ShippingDetails and MerchantReturnPolicy schema on every PDP. If you're on Shopify, audit what Agentic Storefronts is publishing — it uses your existing feed.
  3. Participate in the rails. Get onto Shopify Agentic Storefronts if eligible; evaluate PayPal's agentic commerce programme; ensure your Merchant Center feed is alive and your GMC returns and shipping data is accurate.
  4. Publish the content agents cite. Specific, dated, named-example content outperforms generic "ultimate guide" pages — the agents quote authority, not fluff.
  5. Instrument AI-sourced traffic. Tag inbound referrers from ChatGPT, Perplexity and Copilot. You can't improve what you can't see, and GA4 isn't built for this by default.

What this means for UK marketers and retailers

The traditional marketing stack was built for a world where you earned a click, landed on your site, and optimised the funnel. Agentic commerce collapses that — the "click" is a citation inside an LLM, the "landing page" is an answer sentence, and the "funnel" is a checkout the agent can complete on your behalf.

The teams winning this are not bigger. They are faster at three things: producing the content that gets cited, maintaining the structured data that gets read, and iterating the offer the agent recommends. You can't do any of that at the tempo of quarterly campaigns, agency retainers and siloed tools — which is exactly the problem Anjin was built to solve.

Anjin: the Marketing Operating System for the AI-shopping era

Anjin is the Marketing Operating System. It is the command layer a marketing team uses to plan, produce, publish and measure content across every surface an AI agent might read — from PDP schema and review content to blog posts, social, landing pages and the ad copy that still buys the residual clicks. One workspace, one memory of your brand, one place to ship the assets an agent needs in order to recommend you.

For UK retailers and brands staring down Shopify Agentic Storefronts, PayPal Instant Buy, John Lewis on Gemini and Tesco on Adobe, that single layer is the difference between being cited by the agent and being a footnote on page 3 — which, incidentally, is exactly where this post used to sit before we rewrote it.

The £888 Lifetime License — Offer Closing Soon

Lifetime access to Anjin for a one-time payment of £888. Not a subscription. Not a seat. Not a trial. One payment, unlimited use, for as long as Anjin exists.

The average marketing team spends £888 in about three working days on tooling, freelancers and coordination software. You're buying the platform that replaces most of it — once.

This price will not be offered again once we close our early-access cohort.

Claim your £888 Anjin lifetime license →

Founders, agency owners and in-house marketers — this is how you run marketing at AI speed without the team, the burn, or another year of waiting.

Sources: Shopify newsroom, ModernRetail, PayPal newsroom, Retail Rewired (Debenhams-PayPal), Retail Rewired (John Lewis-Commercetools), John Lewis Partnership, The Grocer, Retail Gazette, CXToday, Fashionista, WWD, Mordor Intelligence, MetaRouter

Continue reading