On July 6, 2026, Salesforce made its biggest bet yet that AI agents — not chatbots, not search boxes — will decide who wins the next holiday shopping season. Agentforce Commerce, the company’s suite of transactional AI agents for retail and B2B selling, is now generally available, arriving with native integration into ChatGPT and plans to reach Google Search’s AI Mode and the Gemini app later this summer. The timing is not an accident: Salesforce is positioning this release squarely ahead of peak 2026 holiday shopping, and the company is citing data showing AI-assistant-driven traffic to retail sites grew 119% year-over-year in the first half of 2025.

Key Takeaways

• Agentforce Commerce launched three purpose-built agents — Shopper Agent, Buyer Agent, and Merchant Agent — that are generally available as of the July 6, 2026 announcement.
• Shopper Agent handles B2C conversations from product discovery through checkout and into post-purchase service, on a brand’s own storefront.
• Buyer Agent moves B2B procurement into WhatsApp and SMS, supporting SKU confirmation by image, contract pricing, and order completion without a portal login.
• Merchant Agent gives commerce and merchandising teams a plain-language interface for catalog management, product sorting, and trend response.
• Native ChatGPT integration is live now; Google Search AI Mode and the Gemini app integrations are slated for later this summer.
• Salesforce reports that retailers running their own shopper agents grew sales roughly 59% faster than retailers still on the sidelines during the 2025 holiday season.
• A new Agentic Commerce Search capability, built on Cimulate technology, is showing a 13% lift in conversion rate and a 17% jump in add-to-cart rate in early deployments.
• Named early adopters include PacSun and Iceland Foods, both of which have publicly discussed their implementations.

Here’s the sentence that should get every ecommerce and B2B sales leader’s attention: Salesforce says intelligent agents are projected to drive roughly a fifth of all global orders during this year’s Cyber Week. That is not a distant prediction about some future wave of AI adoption — it is a claim about the next few months. Whether or not that specific number holds up, the direction is unmistakable: shopping is starting to happen inside AI agents and AI-mediated search results, not just on storefronts and in native apps, and Salesforce just shipped GA product to compete for that traffic.

What Agentforce Commerce Actually Is

Agentforce Commerce is not a single product — it’s three distinct agents built on top of Salesforce’s existing Commerce Cloud data (catalog, inventory, orders, pricing) rather than bolted on as a separate chatbot layer. That distinction matters more than it sounds. Most first-generation retail chatbots could answer questions but couldn’t reliably check live inventory, honor contract pricing, or actually close a transaction. Salesforce is explicitly positioning these three agents as being wired into operational systems from day one:

Shopper Agent

Shopper Agent runs the B2C conversation end-to-end: it checks live inventory, confirms carrier cutoff times, offers store pickup, and closes the sale in a single conversation — speaking in the brand’s own voice on the brand’s own storefront. Critically, it carries full context and purchase history into any follow-up service interaction, directly attacking the classic pain point of a customer having to repeat their order details to a human agent after an AI handled the sale.

Buyer Agent

Buyer Agent is aimed at B2B sellers and meets procurement buyers where they already work — inside WhatsApp and SMS. It supports SKU confirmation via photo, manages negotiated contract pricing, and completes orders without requiring the buyer to log into a web portal. For distributors, wholesalers, and manufacturers whose customers still place large recurring orders by text message or phone call, this is a direct attempt to formalize that channel instead of forcing a portal migration that buyers resist.

Merchant Agent

Merchant Agent is the back-office counterpart: it lets merchandising and catalog teams organize products, adjust sort order, and respond to demand trends using plain-language instructions rather than manual rule configuration. This is less customer-facing and more about compressing the operational overhead that usually falls on a retailer’s internal commerce team during peak season.

Why This Matters Beyond Salesforce Customers

Even if a business isn’t on Salesforce Commerce Cloud, this release is a signal worth reading carefully. The three integrations Salesforce is racing to ship — ChatGPT (live now), Google Search’s AI Mode, and the Gemini app (both arriving later this summer) — are the same surfaces every other commerce and CRM vendor is trying to reach. When a company the size of Salesforce commits its flagship commerce launch to being discoverable and transactable inside a competitor’s AI product (ChatGPT), that tells you where the industry believes purchase intent is forming.

EVP and GM of Agentforce Commerce Nitin Mangtani framed it directly: “The brands that win will have their Shopper Agent live on their own properties for the 2026 shopping season.” He also drew a distinction worth internalizing — referral traffic will increasingly come from AI platforms and social apps, but Salesforce’s bet is that the actual transaction should still close on a brand’s owned channels: its website, its app, its messaging number, its stores. That’s a meaningfully different strategy than simply selling through a third-party AI marketplace, and it’s the same strategic question every CRM and ecommerce buyer needs to answer for their own business right now: do you want AI agents to bring shoppers to you, or do you want to transact inside someone else’s agent?

Practical Use Cases Across Industries

• Specialty retail: PacSun and Iceland Foods are both named early Agentforce Commerce customers, using shopper-facing agents to handle product discovery and checkout during high-traffic periods without proportionally scaling human support staff.
• Distribution and manufacturing: Buyer Agent’s WhatsApp/SMS-first design fits distributors whose B2B customers place recurring orders by text or phone rather than through a self-service portal — formalizing that workflow instead of fighting it.
• Multi-brand or multi-catalog retailers: Merchant Agent’s plain-language catalog management is aimed at commerce teams juggling frequent seasonal assortment changes across multiple storefronts.
• Any business selling through AI-mediated discovery: The Agentic Commerce Search capability (built on Cimulate technology) is reporting a 13% conversion lift and 17% add-to-cart increase, aimed at businesses whose customers are increasingly starting product research inside an AI assistant rather than a search bar.

Benefits and Real Challenges

The upside case is straightforward: Salesforce’s own data claims retailers running shopper agents grew sales about 59% faster than retailers that didn’t during the 2025 holiday season, and AI-assistant-driven traffic to retail sites is already up 119% year-over-year. If that trend continues, businesses that sit out this shift risk losing visibility in exactly the discovery channels their customers are starting to use.

But there are real implementation challenges worth naming honestly, rather than glossing over in favor of the headline stat:

• Data readiness is the actual bottleneck. An agent that can check “live inventory” and honor “contract pricing” is only as good as the underlying Commerce Cloud, ERP, and pricing data feeding it. Dirty product data or stale inventory sync will produce an agent that confidently gives customers wrong answers — which is worse than no agent at all.
• Pricing is consumption-based, and consumption is hard to forecast. Salesforce’s broader Agentforce platform runs on usage-based pricing — roughly $2 per resolved conversation, or Flex Credits at $500 per 100,000 credits (about $0.10 per agent action). Commerce agents are expected to fall under a similar consumption model rather than a flat per-seat license, which means a viral spike in agent-driven traffic during Cyber Week could show up as a real cost spike, not just a revenue win. Confirm exact packaging with a Salesforce account executive before building a budget.
• Multi-channel handoff needs testing, not assuming. Shopper Agent’s promise of carrying context into service interactions only works if the underlying Service Cloud and Commerce Cloud instances are actually integrated — a common gap in orgs that adopted Commerce Cloud and Service Cloud on separate implementation timelines.
• Being present in ChatGPT changes SEO and content strategy. If purchase discovery shifts toward AI assistants, product content, structured data, and catalog feeds need to be optimized for how an AI agent parses and recommends products — not just how a search engine ranks a page.

Implementation Best Practices

Businesses evaluating Agentforce Commerce, or any comparable agentic commerce tool, should treat this as a data and integration project first and an AI project second:

• Audit product catalog, inventory sync frequency, and pricing rule accuracy before enabling a customer-facing transactional agent — not after.
• Start with a single high-volume, well-understood use case (for example, order status and reorder for existing customers) rather than launching a fully open-ended shopper agent on day one.
• Model consumption-based costs against a realistic peak-traffic scenario, not average daily volume, since Cyber Week-style spikes are exactly when these agents are designed to do the most work.
• Keep a clear human escalation path; an agent that closes a sale but can’t gracefully hand off a complex return or dispute will damage trust faster than it built it.
• Treat the ChatGPT/Gemini/Google AI Mode integrations as a distribution and content strategy decision, involving marketing and SEO stakeholders, not just an IT or commerce team decision.

CRM Experts Online’s Perspective

We implement Salesforce, Zoho, and NetSuite environments for clients who are asking a version of the same question this release forces: should we be building our own AI shopping experience, or waiting to see how this shakes out? Our honest answer is that the waiting period is getting shorter than most leadership teams assume. A 119% jump in AI-assistant-driven retail traffic isn’t a rounding error, and Salesforce timing this GA squarely ahead of Cyber Week is a deliberate signal to its retail customer base to move now, not in Q1 next year.

That said, we’ve seen enough Commerce Cloud implementations to know where these projects actually go sideways — and it’s almost never the AI model. It’s inventory feeds that are twelve hours stale, pricing rules that live in three different systems, and service teams that have no visibility into what a shopper agent promised a customer five minutes before the call came in. Before any client of ours turns on a customer-facing transactional agent, we insist on a data and integration audit first — because an agent that confidently tells a customer an out-of-stock item is available does more brand damage in one bad interaction than the productivity gains recover in a month. If you’re evaluating Agentforce Commerce, or wondering whether your current Commerce Cloud, Zoho, or NetSuite setup is even ready for an agent like this to sit on top of it, that’s exactly the conversation we have with clients before we touch a configuration screen.

FAQ

Is Agentforce Commerce only available to existing Salesforce Commerce Cloud customers? Yes, based on available reporting, Agentforce Commerce is built on top of Salesforce Commerce Cloud data and infrastructure, so it requires an existing or new Commerce Cloud deployment rather than functioning as a standalone add-on.

How much does Agentforce Commerce cost? Salesforce has not published Commerce-agent-specific pricing publicly. The broader Agentforce platform uses consumption-based pricing — around $2 per resolved conversation, or Flex Credits at $500 per 100,000 credits (roughly $0.10 per action) — and early indications suggest Commerce agents follow a similar usage-based structure. Confirm exact packaging directly with Salesforce before budgeting.

What is the difference between Shopper Agent and a standard AI chatbot? Salesforce’s stated distinction is that Shopper Agent is natively wired into live catalog, inventory, and order systems from day one and can actually complete a transaction end-to-end, rather than answering questions and then routing the customer elsewhere to check out.

Can Buyer Agent replace our existing B2B customer portal? Not necessarily replace, but it offers an alternative channel — WhatsApp and SMS — for buyers who prefer not to log into a portal, which can be valuable for customers who currently place orders by phone or text with a sales rep.

Does the ChatGPT integration mean our products will show up automatically in ChatGPT shopping results? The native integration allows brands to make their catalog and checkout available through ChatGPT, but appearing in relevant results still depends on product data quality, structured feeds, and how the integration is configured — it is not automatic exposure.

When do the Google Search AI Mode and Gemini app integrations become available? Salesforce has stated these are arriving later this summer 2026, after the initial ChatGPT integration, which is live now.

Is this only relevant to large enterprise retailers? The named early customers (PacSun, Iceland Foods) are established retail brands, but the underlying trend — AI-mediated shopping discovery — affects mid-market and B2B sellers as well, particularly those who already rely on Commerce Cloud or are evaluating a commerce platform migration.

What happens if the AI agent gives a customer wrong information, like incorrect stock availability? This is a genuine risk that depends heavily on inventory sync accuracy and integration quality between Commerce Cloud and backend systems — it is the single most common failure mode in agentic commerce deployments and should be tested extensively before go-live.

Conclusion

Agentforce Commerce is a concrete, dated signal, not speculative AI hype: Salesforce shipped three transactional agents to general availability on July 6, 2026, timed deliberately ahead of peak holiday shopping, backed by real customer names and real (if vendor-reported) performance stats. Whether your business runs on Salesforce Commerce Cloud, Zoho, NetSuite, or another platform entirely, the underlying question is the same one every commerce and B2B sales leader needs an answer to before Cyber Week: is your product data, inventory sync, and channel strategy actually ready for a customer to complete a transaction inside an AI agent? If you’re not sure, that’s worth finding out now rather than during peak season. CRM Experts Online helps businesses audit their commerce and CRM data readiness and implement agentic commerce tools the right way the first time — schedule a consultation with our team before you commit budget to a Q4 rollout.

Further Reading

• Salesforce: As AI Agents Transform Commerce, Salesforce Unleashes Its Biggest Agentforce Commerce Release Yet
• CX Today: Salesforce Makes Agentforce Commerce Generally Available Ahead of Peak Season
• MarTech Notes: Salesforce’s Agentforce Commerce GA Lands Ahead of Peak Season
• Salesforce: AI Assistants Grow Retail Traffic Volumes by 119%
• Futurum Group: Salesforce’s Agentforce Commerce Pushes Agentic AI From Hype to Retail Revenue Reality
• Salesforce: Agentforce Pricing