Accelerating Agentic Commerce: Keys for CPGs
As it relates to innovation within the retail industry, retailers are the fast movers and CPGs are often the followers. In the case of agentic commerce, this trend seems to be particularly true.
Notably, Walmart recently made a splash by announcing its partnership with OpenAI, while Boston Consulting Group reported that only 10% of CPGs have successfully integrated AI agents to support their business teams and workflows. As brands look to catch up with agentic commerce, the first step is to ensure they have the right technical foundation beneath the AI.
Research from a global report by the MACH Alliance finds that more than half of executives believe integration and compatibility issues are a primary cause of why AI projects fail. Nearly 4 in 10 claim legacy system limitations halt AI success.
Gartner’s 2026 predictions claim that over the next two years, generative AI and agentic AI will create the first true challenge to mainstream productivity tools in three decades, adding that companies that create composable architectures will “establish a significant competitive moat.” The research company says that within the next three years, 90% of B2B buying will be AI agent intermediated, pushing over $15 trillion of B2B spend through AI agent exchanges.
Looking beyond the numbers, how can brands educate themselves to build a composable infrastructure that supports agentic commerce, managing inventory predictions, personalizing consumer engagement and optimizing business decisions?
What CPGs Can Do With Agentic AI
The future-forward vision of how CPGs use agentic commerce is very different than how brands sell and engage with consumers today.
AI enables CPGs and retailers to autonomously interact with one another through agents. Brands are able to use the tech to autonomously negotiate pricing and delivery schedules with retailer partners, predict product demand, and adjust demand and production in real time.
However, to reach this level of agentic execution, brands need an ecosystem in which diverse AI agents can communicate, transact and collaborate across departments. All teams need to be connected by unified data streams, fostering an interoperable environment where AI agents work with other AI agents, including those of retailers, logistics providers and payment processors.
To power the future, brands need to ensure they’re building a flexible, interoperable architecture to support the promise of agentic commerce.
Why Composable Systems Will Outperform Monolithic Ones
Fundamentally, what holds back legacy technologies and monolithic platforms from easily powering this agentic future is that these platforms were originally designed for a world of human decision-making and manual processes. When CPGs attempt to layer in agentic AI onto these legacy systems, they often encounter immediate friction.
Monolithic systems also force brands to build AI agents in silos, such as by producing an autonomous purchasing agent that can’t talk to an inventory agent. This fragmentation defeats the core value of agentic commerce: a synergistic intelligence working across agents and collaborating across different business functions.
Ensuring that every part of the business, from supply chain to marketing to IT, is in sync with agentic commerce can take time. This is why CPGs can’t take a “polar plunge” approach toward implementing agentic commerce. Brands need to properly implement AI, continually test solutions, add/replace partners, and leverage the advantages of a flexible, composable architecture to make sure they’re launching autonomous AI agents that know their business.
Interoperability and Composability Build a Path Forward
Composable architectures naturally support the interaction patterns that autonomous agents require. API-first design simplifies agent integration and enables real-time data exchange. Cloud-native infrastructure provides the scalability AI agents demand, powering the ability to spin up resources during peak processing and scale down during quiet periods. Microservices enable the modularity that allows different agents to evolve independently without disrupting the broader system.
These aren't theoretical benefits; they are the foundation for creating an ecosystem that enhances agentic commerce. Brands that avoid single-vendor, monolithic architectures will develop agentic AI without silos and create common data models and communication standards to fuel AI agents.
As Walmart pushes its relationship with OpenAI and more CPGs add agentic commerce, the technology may revamp how brands operate, compete and serve their consumers. The brands that are ready will first implement a technical foundation that’s open, composable and designed for agentic commerce.
Holly Hall is managing director of MACH Alliance, a not-for-profit industry body that advocates for open, best-of-breed technology ecosystems.
