Salesforce moved Agentforce Commerce into general availability on June 29, 2026, releasing three production-ready agents — Shopper, Buyer, and Merchant — along with a native ChatGPT integration that places the platform inside the consumer AI discovery channels retailers have been racing to reach before peak season.
Three Agents Built Around Three Commerce Jobs
Agentforce Commerce GA is organized around the distinct workflows that move revenue: discovery and purchase on brand-owned storefronts, B2B procurement across messaging channels, and back-office merchandising operations. Each agent targets a specific point of friction and is designed to operate without requiring a human handoff for routine transactions.
Shopper Agent: Closing the Conversational Loop on the Storefront
The Shopper Agent operates directly on retailer-owned properties, handling questions that have historically bounced between search, product pages, and customer service queues. It checks live inventory, confirms carrier cutoffs, surfaces store-pickup availability, and closes the sale within a single conversation.
Advertisement
300 × 250
The business case for keeping that channel on owned properties is substantial. Retailers who deployed their own shopper agents grew sales 59 percent faster than those who had not, according to Salesforce data released with the announcement. AI-referred traffic converts at eight times the rate of social referrals.
Buyer Agent: B2B Orders Without the Portal
The Buyer Agent addresses procurement workflows that have historically demanded portal logins, ERP integrations, or sales rep availability. It operates over WhatsApp and SMS, supports image-based SKU confirmation — a buyer photographs a component and receives price and availability instantly — and applies contract pricing to complete orders without any portal interaction.
For sales teams managing high-SKU catalogs or geographically distributed buyers, the Buyer Agent directly reduces the friction that pushes purchasing toward non-preferred channels or competitors with simpler ordering experiences.
Merchant Agent: Removing the Merchandising Bottleneck
The Merchant Agent handles the operational work that consumes merchandising bandwidth: catalog updates, promotion configuration, and back-office administration. By automating routine tasks, it frees merchandisers to concentrate on assortment strategy and pricing decisions rather than platform maintenance.
Storefront Next and Agentic Commerce Search
Two additional infrastructure pieces reach general availability alongside the core agents. Storefront Next — designed to deliver enterprise capability with out-of-the-box simplicity — is now the default deployment surface for new commerce implementations. Luke Barber, Head of Ecommerce Technology at Iceland Foods, described the result: “Storefront Next is exactly where we want to be. It’s fast to run, easy to adopt, and already delivering better than anticipated progress.”
Agentic Commerce Search, built on technology from the Cimulate acquisition, replaces keyword matching with intent-aware product discovery that reads shopper context across browsing and buying signals. A shopper researching a product category in one session sees contextually relevant recommendations in the next, without requiring explicit preference re-entry.
ChatGPT Integration and the AI Distribution Layer
The native ChatGPT integration — live with the GA release — is the element commerce leaders will find most significant for near-term strategy. Google Search AI Mode and the Gemini app integrations are scheduled to follow later in summer 2026.
The distribution logic differs from earlier social commerce experiments. When a shopper asks ChatGPT about a product and the Shopper Agent responds with a purchase-capable interface, the conversion happens inside the AI platform rather than after a redirect to the retailer site. Shirley Gao, Chief Digital and Information Officer at PacSun, noted the strategic opportunity: “Integration with OpenAI presents a powerful opportunity to extend products into AI platforms like ChatGPT.”
Salesforce is also releasing the Agentic B2C Developer Toolkit — a unified command-line interface, Model Context Protocol server, IDE extension, and advanced agent skills — that allows engineering teams to deploy and configure agents using conversational development workflows rather than traditional custom code paths.
Shared Context Across Marketing, Commerce, and Service
A persistent challenge with point solutions in the commerce stack has been data isolation: marketing systems track ad engagement, commerce platforms track purchase history, service systems track complaint records. Each operates on a partial view of the customer. Agentforce Commerce addresses this through shared context across Salesforce’s marketing, commerce, and service clouds, giving the Shopper Agent a unified customer view rather than a transaction snapshot.
“The brands that win will have their Shopper Agent live on their own properties for the 2026 shopping season,” said Nitin Mangtani, Executive Vice President and General Manager of Agentforce Commerce.
What It Means for the Sales Leader
For sales leaders responsible for commerce revenue — especially those managing hybrid B2B and B2C channels — the Agentforce Commerce GA represents a shift in where sales capacity is allocated, not a reduction in sales team relevance.
The Buyer Agent’s ability to close standard reorder transactions over WhatsApp without portal access or rep involvement removes a category of repetitive work from inside sales queues. That time reallocates toward higher-complexity negotiations and relationship-driven opportunities where human judgment is the differentiator.
The ChatGPT and Gemini integrations change the acquisition funnel calculation. If AI assistants are becoming a primary discovery surface — and conversion data suggests they already are for certain purchase categories — then configuring a Shopper Agent that operates effectively inside those environments becomes a direct revenue decision, not a marketing experiment delegated to the digital team.
Sales leaders evaluating Agentforce Commerce should pressure-test two questions before peak season: First, is the Shopper Agent configured for the specific objections and inventory questions their customers actually raise, or is it running a generic script? Second, does the Merchant Agent have current promotion logic loaded, or will catalog administration delays create a gap between pricing strategy and what the agent actually quotes?
The 2026 peak season window is short. Mangtani’s framing — that winning brands will have the Shopper Agent live on their owned properties — functions as a deployment deadline, not a long-term roadmap suggestion. General availability sets the starting point; the configuration work that determines whether the agent performs starts now.
