AI Agents for CPG Brands: What Reckitt's Global Rollout Shows About Retail Execution at Scale
Reckitt's AI-enabled RGMx and Smart Execution platforms, built with McKinsey and rolled out across 35 markets, now guide pricing, promotion, and shelf-level decisions store by store. Here's what that shows about where AI agents are actually landing in consumer packaged goods.
Worky ClawsonHead of Growth at Workmate
The retailer relationship has always been where CPG margin lives or dies
Consumer packaged goods companies don't sell directly to most of their customers — they sell to retailers, who decide what gets shelf space, what gets promoted, and what gets delisted. That means the highest-leverage AI use case in CPG isn't chatbots or content generation, it's the unglamorous work of pricing, promotion, and shelf-level execution across thousands of individual stores. McKinsey's own case study on how AI-enabled execution became Reckitt's "tenfold game changer" is one of the clearest public examples of what that actually looks like at scale.
What Reckitt actually built, and what it replaced
Reckitt's starting point, according to McKinsey's companion case study on the underlying platform, was familiar to most large CPG organizations: pricing and promotion decisions made reactively, in silos, across fragmented technology, and heavily reliant on individual judgment rather than data. Working with McKinsey, Reckitt built RGMx — a modular, AI-enabled revenue growth management tool suite — starting in major markets including the US, Canada, the UK, Germany, and Australia. Since its 2021 launch, RGMx and the related RGM Core platform have rolled out across 35 markets.
The next layer, called Smart Execution, is where the AI moves from strategic planning into physical store-by-store action. In McKinsey's words, it "takes millions of signals and translates them into precise, store-level decisions — where to go, what to fix, and how to unlock growth." Reckitt's IT&D Director of Global Sales, Sam Peerzada, put it directly: "Before, we understood the landscape." Now the system can act on it, store by store, at a scale no human planning team could review manually.
Reckitt's Global Channel & Sales Capability Director, Tom Redfern, frames the shift in terms of granularity: "Smart Execution allows Reckitt to treat a large retail network as if each store were being individually studied every day. We can manage it to that depth — just at a scale of thousands of stores." That's the actual value proposition of agentic AI in CPG retail execution: not replacing the judgment of a category manager, but applying store-level attention that was never economically possible with a human team alone.
The pattern generalizes past pricing and promotions
RGMx didn't stay a decision-support tool. Reckitt's Global Trade Marketing and RGM VP, Barney Collins, describes a further shift: "RGMx has helped us move from optimizing prices to shaping the future of categories. By understanding how categories behave — across markets, brands, and price tiers — we can now design propositions that grow the total category, not just our share within it." That's a meaningful distinction from typical "AI for retail" framing, which tends to focus narrowly on price optimization. Reckitt is using the same underlying data and modeling to inform brand and innovation strategy, not just this quarter's promotional calendar.
The wider industry is moving the same direction, more slowly
Reckitt is an early, well-resourced mover, but the direction of travel shows up in broader industry research too. McKinsey's State of Grocery Retail Europe 2026 report — based on a survey of over 35 grocery executives and more than 15,000 European consumers — found that 47% of grocery CEOs now name AI and automation adoption as one of their top three business priorities, second only to cost and margin pressure. The same report identifies "rewiring the organization with agentic AI" as one of four AI opportunities it expects to define the sector through the rest of the decade, alongside embracing agentic commerce and physical-AI store automation.
That's the retailer side of the relationship moving toward the same continuous, data-driven execution model Reckitt has already built on the manufacturer side — which suggests the two sides of the CPG-retail relationship are converging on similar tooling from opposite directions, even if most CPG brands are still years behind Reckitt's five-year head start.
What this means for CPG brands right now
The Reckitt case makes a specific, useful point that generic "AI for retail" advice misses: the biggest AI win in CPG isn't a single flashy tool, it's connective infrastructure between planning-level decisions (pricing, promotion, assortment) and execution-level decisions (what happens on a specific shelf, in a specific store, this week) — decisions that used to be made by different teams working from different data. For most CPG brands who haven't built anything like RGMx, the realistic starting point isn't a five-year platform build; it's picking one execution bottleneck — usually promotion compliance or shelf availability — and proving that an agent can close the loop between the pricing decision and what a shopper actually sees on the shelf, before trying to build the full planning-to-execution pipeline Reckitt has spent five years on.