Retailers tell UiPath that AI adoption matters more than capability

Adoption is harder than deploying the technology

UiPath hosted its first retail customer advisory board earlier this month in London, bringing together merchandising, trading and technology leaders from eight retailers. The discussion operated under the Chatham House Rule, so UiPath did not identify the participants or attribute comments to individual companies.

According to UiPath, adoption and change management appeared in every session and came up more often than specific products, platforms or features. Retailers described processes that are 20 or 30 years old, along with teams that have historically worked in isolation. One retailer has a dedicated change management team, while others said employees often focus on what a new tool cannot do rather than what it can.

The company said some retailers have modern planning tools in place, but much of the work still happens in spreadsheets. The advice from the group was to start small, make an early success visible and expand from there. One example involved an internal “Ask HR” agent that was tested before being scaled department by department.

Connected systems do not guarantee useful visibility

UiPath also presented findings from The Reaction Gap, its survey of 500 UK retail leaders conducted earlier this year. Ninenty-eight percent describe their systems and teams as connected during peak trading, but only a quarter have the real-time visibility needed to act on problems as they happen. More than two thirds respond only after an issue has affected commercial performance.

The retailers involved in the advisory board recognized the problem, according to UiPath. Different departments may manipulate the same data in different ways, producing conflicting figures across back-end systems. During peak season, some retailers use war rooms filled with dashboards, but still struggle to identify which three issues require immediate attention.

UiPath said poor or siloed data can block agentic use cases before they begin. One retailer is working to make product data intent-based, with the aim of helping customers find what they need more quickly. The company’s broader message was that speed to the customer depends on data that teams can trust.

Retailers want agents to support, not replace, judgment

In discussions about trading decisions, several retailers described a model in which agents handle routine work while people retain responsibility for decisions that require judgment. Agents could prepare recommendations and clear perhaps 90% of routine decisions, while employees approve the 10% that need human review.

UiPath said this approach depends on systems explaining their recommendations, including the reasoning and risk involved. Teams also need enough data literacy to interpret the dashboards they use. The discussion raised a further concern: if agents remove the manual preparation that helps merchandisers develop judgment, retailers must find other ways to train the next generation.

The company’s survey found that 97% of UK retailers have implemented AI in some form, while 47% are still waiting to see commercial impact. UiPath said returns-prediction models may need time to prove their value, but retailers also need clear metrics for judging whether an investment is worthwhile.

The proposed starting point is overlooked operational work

The advisory board identified a middle band of work below major transformation programs as a practical place to begin. UiPath gave non-stock procure-to-pay as an example. Supplier onboarding can take up to eight weeks, and orders placed before completion can create exceptions later. Teams may spend time chasing approvals, managing amendments and working across information silos.

UiPath said AI could help provide a single source of truth, show what is stuck at each stage and reduce manual purchase order creation. The company also said some non-stock orders may not be needed, creating an opportunity to reduce spending as well as improve processes.

For users and the wider retail market, the message from the session is that AI adoption will depend on contained, measurable improvements that employees can see and use. UiPath plans to bring the advisory board together again in April.