Anthropic is trying to make AI-assisted shopping easier for merchants to build. The Register reports that the company this week published a blueprint and templates for Claude-based shopping agents and merchant agents, aimed at helping e-commerce businesses implement automated purchasing workflows. The core pitch is technical enablement. According to The Register, Anthropic says its blueprint includes the harnesses, design patterns, and guardrails needed to get a commerce agent running in days, with reference implementations for retail, travel, telecom, and ticketing platforms. The company is positioning the work around agents that can help customers search, compare, and potentially buy through natural-language requests. The Register reports that Anthropic’s GitHub repository includes functional shopping and merchant agents that can be built with its Messages API, Agent SDK, or Claude Managed Agents. The shopping-agent blueprint includes connection points for product catalogs, online shopping carts, checkout systems, preference databases, and purchase-history databases. Anthropic’s example use case is a customer asking for gear for a weekend trip with two children, after which the agent would identify suitable products. The Register says Anthropic describes guardrails meant to keep recommendations tied to actual catalog data, constrain prices and products, and avoid manipulative upsell patterns. The harder part is not the code. The Register notes that consumers are not uniformly ready to let AI agents make purchases for them. It cites a Gartner survey finding that 11% of consumers are willing to let AI make purchase decisions. An Accenture survey cited in the same report is more optimistic: 74% of consumers would let an AI agent handle routine tasks, 32% are ready to hand over purchase decisions, and 9% are open to fully autonomous shopping before the technology is fully operational. That gap matters because product discovery and comparison are different from purchase authority. The evidence in the report supports a more modest near-term use case: AI agents that help narrow choices, check catalogs, and prepare carts, while humans retain control over final approval. The Register also places Anthropic’s move inside a broader concern over dynamic pricing. It cites Brookings arguing that agentic AI could worsen dynamic pricing by using behavioral signals to tailor prices to individuals. The report also notes that AI surveillance pricing was discussed at a Senate Judiciary subcommittee hearing last month. One example cited by The Register comes from prepared remarks by Lindsay Owens, president and CEO of Groundwork Collaborative. Owens said that half of Walmart app users use Sparky, Walmart’s AI assistant, and that those shoppers spend about 35% more than those who do not. The report does not establish causation from that figure, but it underscores why AI shopping agents will attract scrutiny from consumer advocates and regulators as well as merchants. Who benefits: E-commerce teams already experimenting with Claude may benefit from reference implementations that reduce the engineering work needed to prototype shopping and merchant agents. Consumers may benefit if agents reliably compare products and stay within real catalog and price constraints. Who's exposed: Merchants and AI providers are exposed to trust, compliance, and pricing-scrutiny risks if agents influence purchases or personalize prices in ways customers view as manipulative. Consumers are exposed if purchase-history and behavioral data become inputs for more individualized pricing.