AI agents in AWS Marketplace are reshaping procurement and deployment (with governance cautions)

AWS Marketplace's AI agent catalog has exploded from roughly 1,000 to over 4,000 offerings in the past year, and "AI agents" has jumped from the 64th most searched keyword to the third. For builders shipping AI products or managing enterprise deployments, this shift means agent-powered procurement is moving from novelty to infrastructure. But the same automation that reduces IT overhead also introduces governance gaps that teams need to plan for.

AI agent catalog grows 4x as procurement goes autonomous

According to Matt Yanchyshyn, VP of AWS Marketplace & Partner Services, the marketplace now hosts almost 40,000 applications, with agent offerings rising from about 1,000 to over 4,000 source. The global agentic procurement software market was valued at $1.2 billion in 2025 and is projected to reach nearly $9 billion by 2034, with a 24% CAGR source. Agentic procurement software uses autonomous AI agents to plan, execute, and adapt multi-step procurement tasks without requiring a human to initiate each action source.

What this means for builders: less overhead, more governance

Agent-powered marketplaces are reducing IT overhead by handling discovery, vendor comparison, licensing, entitlement management, auditing, renewals, and even portions of contracting and procurement, according to Bret Greenstein, chief AI officer at West Monroe Partners source. For builders, this means faster deployment cycles and less manual due diligence. But it also raises the stakes for governance. Greenstein warns that when software can be discovered, evaluated, and provisioned faster, enterprises need clearer policies around approved vendors, security, data access, architecture, licensing, and agent authority source.

CharmHealth's AI assistant Ria illustrates how agents can improve the buyer experience. Ria helps prospective buyers evaluate products and streamlines onboarding, collecting higher-quality information for sales teams source. This pattern is relevant for any builder selling through the marketplace: agent-assisted discovery can shorten sales cycles while improving lead quality.

Caveats: agent search coverage and the need for human review

Shashi Bellamkonda, principal research director at Info-Tech Research Group, cautions that AI agents may only look at less than 70% of available solutions. This is not an AWS issue but a model limitation source. He advises reprompting and adding more details to improve results, and always reviewing output through human eyes. AWS's Yanchyshyn also notes that while medium and lower-level deals are increasingly automated, sophisticated deals with custom pricing and configurations still require experienced sales and procurement teams source.

For AI builders, the growth of agentic procurement in AWS Marketplace is a signal to invest in agent-compatible listing strategies and governance frameworks. The automation is real, but so are the blind spots.