BigCommerce Moves Agentic Commerce Into Authenticated B2B Buying
BigCommerce’s latest MCP update gives AI agents controlled access to shopping lists, approvals and quotes. The real test is whether those agents can act within the buyer’s organizational permissions.

Most commerce agents can help a buyer find a product. The harder test begins after the buyer signs in.
That is where account permissions, approval rules, shared shopping lists, quotes and active carts enter the workflow. BigCommerce took a meaningful step into this territory on September 16 by adding authenticated shopping sessions and new B2B tools to its Model Context Protocol server.
From product discovery to buying action
The release introduces Storefront Session Sync, which connects an agent’s MCP session to a shopper’s existing Stencil storefront session through a short-lived, opaque token. BigCommerce says the buyer’s login state, identity and cart can pass into the agent without exposing browser cookies. The storefront can also remain synchronized if the agent creates a new cart.
For B2B buyers, the platform now offers 12 shopping-list tools. These cover actions such as creating and updating lists, adding or removing products, submitting a list for approval, and approving or rejecting it. Five quote tools let authorized buyers request a quote, review details, message a sales representative and generate a checkout link.
The real test is authority
The B2B tools require BigCommerce B2B Edition and are made available according to the buyer’s permissions. That distinction matters. A user may be able to prepare a shopping list but not approve it. Another buyer may be authorized to request a quote without being able to accept its commercial terms.
This is the point where agentic B2B commerce becomes more consequential. Authentication confirms who is present. It does not, on its own, establish everything that person is allowed to do across different companies, branches, budgets or purchasing roles.
For manufacturers and distributors, the value will depend on whether the agent preserves this context throughout the transaction. Commerce teams should test:
whether the agent receives exactly the permissions assigned to the buyer;
whether approval workflows remain intact;
whether products, quantities and messages stay aligned throughout a quote;
what happens when a session expires, an account changes or a cart already exists; and
whether the business can reconstruct who initiated, performed and approved each action.
Account-specific product access, negotiated pricing and inventory rules also need direct validation. BigCommerce’s announcement documents identity, carts, permissions, shopping lists and quotes, but does not provide production evidence showing how every connected commercial rule will perform in practice.
The release is still an important market signal. BigCommerce is moving agents beyond product recommendations and cart creation into recognizable B2B buying workflows. The next proof point will be whether customers can reduce the effort involved in repeat purchasing and quote preparation without increasing pricing errors, unauthorized actions or manual cleanup.
The question for B2B commerce teams is no longer simply, “Can the agent find the right product?”
It is, “Can the agent act for this buyer, inside this company’s rules, and prove that it did?”
Source: BigCommerce developer changelog, September 16, 2026
Editorial note: Product capabilities are based on BigCommerce’s technical announcement. Operational implications and evaluation recommendations are B2BEA analysis. No customer performance data was included in the source.