A shopper agent is a brand-owned AI agent that carries the shopping conversation. From discovery through checkout and into service. In your brand's voice, and increasingly does so inside third-party AI surfaces like ChatGPT, Gemini, and Google AI Mode. The mid-2026 launches made the pattern concrete: Salesforce's Agentforce Commerce reached general availability, and Google rolled out a brand-voice "Business Agent" in Search. The shift is real, but deploying an agent is a new channel, not a new ranking lever. Whether an engine surfaces your product still depends on the same catalog and entity work as before.
This is part of our agentic commerce series. Here's what a shopper agent actually is, what changed this summer, and. The honest part. What it does and doesn't do for your visibility.
What is a shopper agent?
A shopper agent is an AI agent that a brand deploys to handle the customer's shopping journey conversationally, rather than leaving the customer to navigate a site or a generic assistant. It answers product questions, compares options, completes checkout, and handles post-purchase service. Speaking in the brand's voice on the brand's behalf.
The mid-2026 packaging from Salesforce, as reported, splits the idea into three roles:
| Agent | Who it serves | What it does (as described) |
|---|---|---|
| Shopper Agent | Consumers | Carries the conversation "from discovery through checkout and into service," in the brand's voice |
| Buyer Agent | B2B buyers | Orchestrates procurement in the channels buyers use. Reported as WhatsApp and SMS |
| Merchant Agent | Operations teams | Organises catalogs and tunes sort/boost rules from plain-language instructions |
Google's comparable move is the Business Agent: described as a virtual sales associate that answers product questions in a brand's voice directly inside Search. The through-line across vendors: the brand fields its own agent, and that agent is surfaced where shoppers already are.
What changed in mid-2026?
Brand-owned agents moved from preview to generally available, with distribution into the major AI engines as the headline. The dated, attributed picture:
- Salesforce Agentforce Commerce reached general availability in late June 2026, with the Shopper, Buyer, and Merchant agents live. Salesforce set native ChatGPT integration to July 2026, and Google Search (AI Mode) and the Gemini app to follow later in the summer.
- Google expanded its Universal Commerce Protocol and launched agentic shopping tools, including the Business Agent (with named launch partners such as Lowe's, Michael's, Poshmark, and Reebok) and a cross-retailer "Universal Cart," built with commerce platforms including Shopify, Etsy, Wayfair, Target, and Walmart.
Salesforce also reported some eye-catching commerce figures, that AI influenced about 20% of global online sales (~$262 billion) in the past holiday season, and that retailers deploying their own shopper agents grew sales faster than those without. Treat these as vendor-reported and directional: they come from the company selling the product, and the causal claims are not independently verified. The durable fact is the architectural shift, not the marketing numbers: brands are now fielding their own agents inside the engines.
Does deploying a shopper agent make your products more visible?
Not on its own, and this is the distinction that matters. A shopper agent gives your brand a voice and a checkout inside AI surfaces once a shopper is engaging with you. It does not decide whether the engine brings your product into the conversation in the first place. Those are two different layers:
- Selection: whether ChatGPT, Gemini, or AI Mode surfaces and recommends your product from the whole market. Is decided by the factors that make an engine recommend a product: catalog richness, structured data, reviews, entity strength, and corroboration. This is answer-engine optimisation.
- Conversion: how well you answer, compare, and close once you're in the consideration set. Is where a shopper agent helps.
A shopper agent with nothing behind it in the selection layer is a well-spoken salesperson no one is sent to. That's the same lesson as the payment rails: plumbing that activates after the choice doesn't make the choice. The agent is the storefront-in-the-answer; being shown the storefront is still earned upstream.
How do you get your products ready for shopper agents?
Make your catalog legible to machines and your brand trusted by engines. The same agentic-commerce readiness work, now with a clear payoff. In priority order:
- Enrich the catalog with conversational attributes. Beyond price and title, expose answers to common product questions, compatible accessories, and substitutes. The exact attribute types Google added to Merchant Center for conversational surfaces. Prepare your product catalog for AI agents covers the mechanics.
- Write conversational product pages that read like answers, not brochures, so both the engine and the agent can lift a clean, specific claim.
- Make your site an agent-readable endpoint: reachable, server-rendered, structured, and exposed through the shared commerce protocols agents speak.
- Invest in entity strength and corroboration. A strong, corroborated entity is what gets a brand recommended across engines, agent or no agent, and it's the layer no single vendor's tool can sell you.
- Measure per surface. Deploying an agent doesn't tell you whether ChatGPT or AI Mode actually surfaces your products; only measurement does. Track each engine as its own line, over time.
A shopper agent lets your brand speak and sell inside the answer. It does not decide whether the answer includes you. Win the selection layer first, then let the agent close.
The agentic storefront is arriving on every major engine at once, but the brands it lifts will be the ones the engines already choose to surface. Knowing whether AI engines and agents actually find, cite, and recommend your products, as these agents go live, is exactly what Buffy Intel measures. Questions: [email protected].