How the Co-op Group UK Makes AI Work

The Co-operative Group Ltd. provides supermarkets, funeral services, insurance, and personal legal services to over 6 million active member-owners in the UK. It is one of the world’s largest organizations of its kind—a business in which voluntary members share ownership and set policy. The Co-op Group, which is included in Futuriom Enterprise AI Index, supports over 2,300 food stores and 800 funeral homes (it’s become the UK’s largest provider of funeral services) and employs over 19,500 workers.

AI is an ingredient in the ongoing progress of the Co-op Group in serving its members with growing revenues. The organization now features several AI use cases with more in trial. Futuriom caught up with Scott Robertson, Principal Enterprise Architect-Foundation Technology, to explore how the Co-op Group’s adopted AI and how it’s unfolding it in day-to-day operations. (You may recall Scott from our Cloud Tracker Pro Premium Q&A interview last year.) Read on!

The Co-op Evaluates the Value of AI

Like many enterprises, the Co-op Group has taken a measured approach to AI adoption. There is a sense that “just because you can doesn’t mean you should” when it comes to deploying the technology, said Robertson. That said, since the Co-op already had a data policy and data engineers overseeing business intelligence (BI) and analytics functions, it was a logical step to create an AI policy when the technology gained momentum. The Co-op also assembled an in-house team of 15 or so data engineers (tasked with building and training models); data scientists (who deal with adjusting the models based on output); plus line-of-business personnel and enterprise technology architects, including Robertson.

The team outlined its goals early on. AI would be used to streamline work, not replace personnel. “Being more productive, it's not about replacing staff, it's about automating or simplifying the mundane tasks or the repetitive tasks that folk have to do to give them more hours back in their day,” Robertson said. “It's not about, oh, I could replace you with a clever piece of technology here. We're definitely not about that.”

A second and just-as-important strategy was that any AI use had to address something that customers actually wanted, not something that would simply be flashy and fun to deploy and may just happen to catch on. In that sense, the Co-op’s approach is “considered,” in Robertson’s view. Some AI applications were trialed but not released because they wouldn’t have significant customer impact. The ones that have been chosen for production have proved to increase productivity and sales.

Scott Robertson. Photo by Mary Jander via Teams

Replacing Manual Probate Tasks

One of the first AI applications the team created was designed to streamline the task of probating wills in the Co-op’s Legal Services division. An application built using Microsoft’s OpenAI version of ChatGPT 4 processes an image of a will needing probate and converts that to text. (This was needed since the handwriting of many older wills, which are often required during probate, was nearly indecipherable to modern eyes.) Once the text is delivered, the application parses the content and builds a hierarchy within and among various documents, indicating which ones take precedence legally.

The team spent about six months training the probate model, which has been in production since the start of 2025. “I think it’s about ninety-nine percent accurate now in terms of how it would correlate with what human beings are doing. So it's freed up the legal staff to actually add value rather than just trolling through documentation,” Robertson said.

Photo by Mary Jander via Teams

More Uses and Plans for AI

Given the Co-op’s focus on retail, the team is exploring a range of AI possibilities. One includes augmenting the chatbots used by clerks in the store, who typically are equipped with headsets and tablets with text-to-speech capabilities. Where now a typed query can deliver a range of hierarchical answers to questions on policy and procedures, the team envisions AI that would allow a clerk to ask a question via headset to an application that converts speech to text, processes instructions, and returns audio text-to-speech responses. That’s currently in trial.

Yet another potential application involves auctioning ad space on a store display to attract customers as they enter the store. “You can sell real estate on digital marketing screens… there are real, live auctions for advertising to a particular demographic,” Robertson said. “A fifty-year-old white man walks in. By the time I’ve walked to the screen, vendors have auctioned for the right to display ads targeted specifically to me in real time…. Then you’re directed to an aisle with an electronic shelf that’s flashing because there’s a proximity sensor that knows where my phone is in the store. The shelf with the offer on is now flashing in front of me.”

The Co-op already engages AI to personalize offers tagged to a membership wallet app, but real-time auctioning of demographic information is something that may be in the organization’s future. “That’s something for the data scientists to work on,” Robertson said. “There’s a little bit of AI that’s embedded in that, but it’s more about heuristic data mining as to what offers I would get.”

AI Adds Productivity Gains

The Co-op Group uses Microsoft Copilot across Teams, SharePoint, and Office applications to improve productivity among office workers. But there’s been a catch: To protect data and ensure accuracy, Copilot has access to only the Co-op’s own data. The organization set a policy against any AI application using data from any open, publicly available source or model.

“We are now at the stage where data is being generated by artificial intelligence,” Robertson said. “I’m really, really nervous about the broader implications of AI… Too many people, I think, are trusting that it gives them the right answer. Now I do use it personally, but I’m quite careful about what questions I ask. I don’t ask for answers, I ask for techniques about how I might best to and find those answers.”

Futuriom Take: The Co-op Group’s use of AI follows the measured pace the organization’s used to adopt technologies up to now. The membership has prioritized serving actual internal and external customer needs over introducing applications for which a market is speculative.