Anthropic on Tuesday launched Claude for Commerce Agents, a blueprint that gives engineering teams the harnesses, patterns, and guardrails needed to build shopping and merchant agents on top of Claude. The company said the package includes working reference implementations and a Claude Code plugin, and is available immediately, just in time for what Anthropic described as holiday season planning.

Retailers are already running shopping agents on Claude For more context on this story, see our ongoing artificial intelligence updates.

According to Anthropic's announcement, many of the world's largest retailers, marketplaces, e-commerce platforms, and travel companies already use Claude to build agents that let consumers search, compare, and buy in plain language. The company named Shopify and Priceline among the enterprise customers with such agents in production.

The business results Anthropic attributed to these deployments are notable, though they come from the company's own blog rather than an independent audit: retailers running shopping agents on Claude have seen carts up to 35 percent larger, and shoppers have been 60 percent more likely to complete a purchase.

What is in the blueprint

The repository contains complete, working implementations of two agents: a shopping agent that faces the customer, and a merchant agent that supports the people running the store. Both can be built using the Messages API, the Agent SDK, or Claude Managed Agents, which is currently in beta. Teams can run a self-guided demo before writing any code, then use Claude Code to customize the agents to their own catalogs, policies, and brand.

The shopping agent is designed to live inside a retailer's app or website. Anthropic says the blueprint includes integration points for catalog, cart, checkout, customer preferences, and order history, while leaving payment handling to the merchant — whether that is an existing checkout or an agentic payments provider. In the company's example, a customer can say "I need a tent, sleeping bag, and stove for a weekend trip with two kids," and the agent assembles the right items, shows products and comparisons in the conversation, builds the cart, and hands it to checkout. The same conversation can handle customer service questions such as order status, returns, and refund policy.

Guardrails against manipulative upselling

One detail that stands out in the announcement is the explicit treatment of guardrails. Anthropic says the shopping agent's controls are designed to constrain prices and products to actual catalog data, and to avoid manipulative upsell patterns. Inside the repository, these capabilities ship as skills and tools for catalog search, multi-item planning, deep research, personalization, customer care, and in-conversation UI.

The merchant agent, by contrast, is aimed at store operators. Anthropic says it can answer questions about sales performance, track inventory and proactively flag problems — for instance, an item about to sell out before a promotion starts — recommend pricing and promotions based on the store's own sales history, and draft marketing campaigns. When the agent proactively suggests a change, a person approves it before anything goes live.

Payments giants are partnering on the effort

Anthropic is not doing this alone. The company said the blueprint works with solutions and ecosystem partners including Accenture, Mastercard, and Visa, which are working with Anthropic to bring the blueprints to their clients and merchant communities.

The partnership quotes in the announcement underline how seriously the payments industry is taking agentic commerce. Jack Forestell, Visa's chief product and strategy officer, said AI will fundamentally reshape commerce but that trust must remain at the center of every transaction, adding that merchants want more control over how AI engages their customers. Sherri Haymond, Mastercard's executive vice president and global head of digital commercialization, said trust is even more critical in the agentic era and described the collaboration as helping connect consumers, merchants, and AI agents securely and at scale.

Accenture, in its quoted remarks, cited its own research finding that 85 percent of consumers are now open to collaborating with an AI agent, and that nearly three in four would trust a personal AI agent more than their best friend to make a purchase on their behalf. Those figures come from Accenture's survey work, not from observed shopping behavior, but they signal how far consumer acceptance of agentic shopping has moved in the industry's eyes.

Deployment flexibility for enterprises

Anthropic said the code deploys wherever teams already build with Claude: the Claude API, Amazon Bedrock, Microsoft Foundry, or Google Cloud Vertex AI. Live demos are available for each vertical the blueprint targets — retail, travel, telecom, and ticketing — along with an engineering deep-dive explaining how the reference implementation was built.

Why it matters

The launch is a sign that agentic commerce is moving from pilot projects to packaged products. Rather than selling a closed shopping assistant, Anthropic is selling the scaffolding — the harnesses, guardrails, and reference code — that lets retailers deploy their own agents under their own brand, with humans keeping approval rights over consequential merchant-side changes.

The timing is also strategic. With holiday season planning underway, retailers face a narrow window to retool their storefronts before the year's biggest shopping months. If Anthropic's reported conversion numbers hold up across more merchants, the pressure on retailers to offer a competent shopping agent — and on rival AI providers to ship comparable commerce tooling — will only increase.

Developers can find the blueprint, the reference implementations, and the Claude Code plugin through Anthropic's announcement page, and the company said the package is available globally today.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →