What was announced
The Very Group, the UK online retailer behind Very and Littlewoods, has selected UiPath to deliver an agentic AI pricing solution under a three-year partnership announced in mid-July. UiPath won the work through a competitive pitch, and the system will bring faster, more transparent pricing decisions across the group's retail brands. The scope is deliberately broad rather than a narrow pilot on one category, which signals that The Very Group is treating pricing as a core process to rebuild rather than a feature to experiment with on the margins of the catalog.
The prize is control over a very large catalog. Sam Wright, chief customer and commercial officer at The Very Group, said the business carries a range of over 200,000 products and that pricing is one of the most powerful levers in retail. At that scale, repricing by hand is slow and inconsistent, and every day a price sits wrong is either margin left on the table or volume lost to a sharper competitor. The three-year term also tells you this is a commitment to operationalize the technology rather than a proof of concept that can be quietly shelved after a quarter.
Why pricing is the right target for agents
Pricing is an appealing first home for agentic AI because it is high frequency, data rich and directly tied to money. A catalog of 200,000 items facing shifting demand, competitor moves and stock positions generates far more pricing decisions than a human team can make well each day. An agent that can read those signals and act continuously changes the cadence from periodic reviews to near-constant adjustment, and it does so on the exact axis where retail margins are won or lost in a competitive online market.
The Very Group frames the payoff in concrete operating terms: optimized gross margin, better stock management and improved pricing agility in an increasingly competitive retail environment. Those are the metrics a chief commercial officer is measured on, which makes the business case legible to the finance side of the house rather than resting on a vague promise of AI efficiency. By moving routine repricing to an agent, the retailer also frees its analysts to work on the harder, more strategic pricing questions rather than grinding through spreadsheets of individual line items that add little value to reprice by hand.
Explainability is part of the deal
The detail that should catch a technology leader's eye is that UiPath is selling explainability alongside autonomy. The solution enhances The Very Group's existing data-led pricing with clear AI explainability, plus campaign simulation, optimization and scenario planning. In pricing, a black box is a liability, because merchants, finance and sometimes regulators all need to understand why a number moved. An agent that can justify its recommendations is one a merchandising team will actually let run, and one an auditor can sign off on when a promotional price draws scrutiny.
Campaign simulation and scenario planning matter for adoption as much as accuracy. They let a merchandising team test a promotional plan or a margin target against the agent before committing, which builds the trust needed to let the system act on its own. Catherine Frame, director of retail solutions at UiPath, said the partnership demonstrates how agentic AI can fundamentally reshape retail pricing. Reshaping it in production, though, depends on humans being able to see and override the logic, so the simulation layer is less a nice-to-have than the mechanism that gets the agent past internal risk committees and into live decisions.
From pilot to production
Most agentic AI in retail is still stuck in pilots, so a three-year commitment across a full catalog is a notably concrete step. The Very Group is not testing the idea on a handful of SKUs; it is rebuilding a core commercial process around it. The partnership also aims to align the operating model and strip out manual steps, so the change touches how pricing teams work day to day rather than only what tools they open. That operating-model piece is often where these programs stall, because the technology works but the organization around it never adapts to trust the output.
That production ambition raises the bar on governance. Handing margin decisions to software across 200,000 products means the retailer needs monitoring, guardrails and clear escalation paths for when the agent gets a call wrong. The explainability and simulation features are the mechanisms that make that oversight possible, and they are the difference between an agent that stays in production and one that gets switched off after the first bad week of margins. A retailer that cannot answer why the agent cut a price in real time will not leave it running long enough to see the compounding benefit.
What it signals for UiPath
For UiPath, this is as much a survival move as a sales win. The company built its business on robotic process automation, and analysts have openly questioned whether that model gets absorbed by general-purpose AI. UiPath stock closed around $12.15 on July 17, up 1% on the news but still down roughly 26% year to date, and its most recent quarterly revenue grew 17% year over year to $418 million. The market is treating each agentic win as evidence for or against the thesis that UiPath can evolve rather than be replaced.
Winning a marquee agentic pricing deal lets UiPath tell a different story: that it can package AI into an outcome retailers will pay for rather than being displaced by it. Commentators still read the deal as one step rather than a turnaround. For retail buyers, the takeaway is to weigh whether an established automation vendor pivoting to agents offers more production discipline than a younger AI-native rival, and to price in the vendor's own stability when signing a three-year term. A supplier under pressure to prove its model can be an eager partner, but it is worth confirming the roadmap outlasts the current quarter's stock story.
The read for retail leaders
The Very Group deal is a useful template for any retailer weighing where to point agentic AI first. Pricing checks the boxes: it is measurable, it moves margin directly, and it is bounded enough that guardrails and explainability can keep it safe. That combination makes it a lower-risk proving ground than letting agents loose on customer-facing experiences before the governance is ready, and it produces a financial result clean enough to justify the next phase of investment to a skeptical board.
The build-versus-buy lesson is equally clear. The Very Group is buying the agentic layer from UiPath rather than assembling its own, and betting that a vendor with campaign simulation, scenario planning and explainability already built will reach production faster. For technology leaders, the question is not whether to apply agents to pricing but whether to trust an external platform with a core margin lever, and what monitoring you keep in-house to make that trust defensible. Owning the guardrails and the data even while buying the engine is how you keep leverage over a process this central to the business.
