Tesco’s AI Agents Could Soon Do Your Shopping For You

Tesco, the UK’s largest supermarket chain, has spent 30 years learning what millions of people buy. Now it wants AI agents to start acting on that knowledge.

The retail giant accounts for almost 30 percent of the UK grocery market, and through its Clubcard loyalty program it has built an extraordinary picture of British shopping habits, with data said to cover eight out of 10 households. For years, Tesco has used that information to predict what customers might want and personalize the offers they see.

Now it is taking the next step.

Through partnerships with Adobe and other AI companies, Tesco is exploring AI agents that can move beyond predicting customer behavior and actually take action. That includes an AI assistant capable of planning meals, helping customers reduce food waste and automatically building shopping baskets based on their preferences and purchase history.

It raises a much bigger question. Are we approaching a future where AI doesn’t simply recommend what we should buy, but increasingly does the shopping for us?

Tesco And Adobe. What’s The Plan?

In April this year, Tesco announced a strategic partnership with Adobe to create the Tesco x Adobe Innovation lab. This will see the supermarket giant’s data and AI teams working directly alongside Adobe engineers to develop bespoke agentic solutions.

Its technology is built around Adobe’s Firefly Foundry, its genAI platform for personalized, branded creative materials at scale. This is a technological leap forward from Tesco’s previous AI-driven personalization initiatives, which were mainly focused on predictive analytics and recommendation engines.

While those technologies basically provided insights into what someone might buy next based in what they bought before, agents take things further.

Rather than just giving the retailer information, they are capable of taking action. This means reading transactional data and customer behavior signals, generating offers and promotions on the fly, and deciding what way of delivering them to the customer will be most helpful.

One feature currently being tested by staff before it’s rolled out to customers is an AI assistant that will live inside its Clubcard app. It will provide meal planning advice tailored to individual preferences and dietary needs, as well as advice on reducing food waste by using up leftovers at home.

It’s also reported that it will automatically build shopping baskets based on preference and purchase history. This certainly sounds like a step towards the buying capabilities we’ve been told to expect from agents.

The project is new, so no concrete results are available yet. But it sends a clear message that retailers are moving beyond traditional predictive AI and analytics, towards adopting agentic, action-taking models and platforms.

Separately, the supermarket giant has also entered into an agreement with French AI developer Mistral, focused on developing technologies for its internal operations.

According to its press release, this will include developing solutions for content development, document drafting and data analysis, as well as “developing new ways for colleagues to easily access information which they can use to help customers.”

Agentic AI And The Future Of Retail

The bigger prize for AI in retail is easy to see: a future where AI agents shop for us.

They could research products, compare options, build baskets and eventually make purchases on our behalf.

At that point, AI would move from influencing what we buy to actively participating in the buying process.

Tesco isn't there yet, but its current experiments point in that direction. An AI assistant that plans meals or builds a shopping basket may feel like a relatively small step, but each useful interaction gives customers another reason to hand over a little more control.

And that may be the real challenge for autonomous retail.

Tesco already has the data and increasingly has the technology. What it still has to earn is trust.

If customers become comfortable letting AI make the small decisions, the much bigger question is how long it will be before they are willing to let it make the purchase too.