Retail Insight launches AI analytics platform for retailers

Retail Insight has launched DecisionInsight, an AI analytics platform for retailers that lets teams query operational data in natural language.

The platform sits within Retail Insight's wider retail operations offering and is already live with major retailers globally and with 20 suppliers across Europe.

DecisionInsight is designed to address a common retail operations problem: the gap between the volume of data available to businesses and the speed at which staff can turn that data into decisions. An experienced senior analyst can spend up to four hours investigating a single performance question, while the new system is intended to surface answers and the factors behind them in minutes.

Users can ask questions in everyday language and receive a written response supported by charts and underlying figures. The platform draws on a retailer's operational information, including sales, deliveries, promotions, on-shelf availability, inventory, waste and markdowns, labour, forecasting, weather and supplier performance.

The system combines three elements: a retailer's own operational data and semantic model, Retail Insight's data and retail knowledge built over two decades, and current AI models. The semantic layer is designed to reflect how each retailer defines its business, so answers are based on the organisation's own terms and structures.

Answer trail

A central feature of the product is visibility into how an answer is produced. Users can follow the reasoning behind responses, distinguish between established facts and inferred conclusions, and trace figures back to the analysis that generated them.

That approach addresses a persistent concern around the use of AI in business decision-making, particularly in areas such as supply chains, store operations and category management, where an incorrect conclusion can have immediate commercial effects. By showing the analytical path behind a response, Retail Insight is positioning the tool as something closer to an operational assistant than a black-box chatbot.

The platform also frames answers according to the user's role. Retail Insight has built dedicated agents for category owners, senior leadership, supply chain teams, store operations leads and store managers, each designed to present findings and next steps in a way that is relevant to that function.

In addition to responding to ad hoc questions, the system can run analyses on a schedule and distribute them automatically to relevant users. That could include weekly performance updates for category teams or supplier scorecards sent directly to partners, reducing the need for staff to manually compile and circulate recurring reports.

Broader push

The launch adds to a broader push by software providers to bring generative AI tools into operational decision-making rather than limiting them to customer service or content tasks. In retail, that shift is focused on giving frontline and middle-management teams direct access to insights that might previously have required support from analysts or data specialists.

For Retail Insight, the announcement also expands a platform that processes 15% of the world's grocery data across more than 68,000 stores, according to the company. It counts Kroger, Sprouts Farmers Market and Marks & Spencer among the retailers it has worked with on store operations issues such as waste, stock records and shelf availability.

DecisionInsight is available through a web application, a mobile app and as an integration for retailers running their own AI agents. The integration uses the Model Context Protocol, allowing retailers to connect the system to internal tools already used by their teams.

Alex Considine Tong outlined the rationale for the launch in comments accompanying the announcement.

"Retailers have no shortage of data. The difficulty has always been getting to answers quickly enough to act on them," said Alex Considine Tong, chief product officer at Retail Insight.

"AI can analyse vast volumes of information at extraordinary speed, but without context and a deep understanding of that data, it can just as easily lead you to the wrong conclusion. Very few companies can bring the context and the technology to the same problem, and twenty years in retail technology and data is what lets us do it," said Tong.

He said the company sees the product as a way to widen access to operational analysis across different levels of a retail business.

"What excites us is how ordinary this makes it feel," said Tong. "A store manager, a category owner and a director can all ask the same question in their own words and get an answer they can act on before the day is out. That is a very different way of working with retail data, and it allows retailers' analytics teams to operate where they belong, on the problems only they can solve."