Anthropic Releases Free Blueprint for Building Shopping Agents with Claude

By Carlos Montiel | Enterprise AI Specialist
Leer en español →
Published: 2026-09-12 | By: Carlos Montiel | Reading time: ~4 minutes

On September 2, 2026, Anthropic published Claude Commerce Agents, a free reference repository (github.com/anthropics/commerce-agents) with two already-built agents — one for the shopper and one for the merchant — designed so retailers, airlines, telecoms, and entertainment platforms can deploy agentic commerce without starting from scratch.

Two Agents: One for the Shopper, One for the Merchant

The shopper agent lives inside a merchant's app or website and helps customers search for products, compare options, build multi-item purchases, receive personalized recommendations, and get answers about their orders. The merchant agent, by contrast, is built for store staff: it analyzes sales, monitors inventory, suggests pricing, and drafts marketing campaigns — but any real change it proposes requires human approval before it executes.

Available on Claude API, Bedrock, Vertex AI, and Microsoft Foundry

The blueprint ships with reference harnesses, patterns, and guardrails, and can be deployed via Claude's Messages API, the Agent SDK, Claude Managed Agents, or directly on Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Foundry — meaning it doesn't lock a company into a single cloud provider to adopt it.

Consumer Trust, Not Technology, Is the Real Bottleneck

Anthropic cites data showing the gap between technical capability and consumer willingness: according to a Gartner survey, only 11% of consumers are willing to let an AI agent make purchases on their behalf. Accenture found that 74% would delegate routine tasks, while only 32% are ready for purchase decisions and just 9% would accept fully autonomous purchases. As a real-world impact reference, Anthropic notes that users of Walmart's "Sparky" AI assistant spend roughly 35% more than non-users.

Risks the blueprint itself acknowledges: Anthropic warns about the risk of discriminatory dynamic pricing if an agent adjusts offers based on each shopper's individual history, and about a liability gap around fraud — there is currently no clear consensus on who is responsible when an AI agent makes a purchase the consumer disputes as unauthorized.
What this means for a company that uses or buys AI: this release is the technical counterpart to the Know-Your-Agent framework from Visa, Mastercard, and Ant International announced days earlier — while payment networks work on verifying who a shopping agent is, Anthropic is already giving away the code to build one. If your company sells online in Guatemala or Latin America, this is a good moment to audit whether your e-commerce platform can expose an API that a shopping agent can safely consume, instead of assuming all agent traffic is suspicious traffic.
Carlos Montiel
Enterprise AI Solutions Architect
Specialist in LLMs, Agents and Orchestration
guatemalia.com/en/#contacto · info@guatemalia.com

Need to implement AI at your company?

Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, Agents, RAG, and orchestrators for companies across Guatemala and Latin America. Reach out for a consultation.

Contact Carlos Montiel

info@guatemalia.com