TLDR;
Google introduced Business Agent for selected retailers in Search. It answers product questions in the retailer's voice before customers reach its website. Brands need accurate specifications and policies behind those answers, plus a clear route to the right product and purchase page.
What happened
Google introduced Business Agent, a branded conversational experience in Search that answers product questions in a retailer’s voice. Initial availability involved selected retailers and eligibility requirements, with future capabilities distinguished from those available at launch.
Why it matters
Answering in a retailer's voice is valuable only if the information is reliable. Incorrect compatibility advice, missing restrictions or outdated delivery promises can damage the brand before a customer reaches its website. The enterprise opportunity is a useful product service close to demand, with accountable information and a clear route to purchase.
A retailer's Business Agent can answer product questions closer to the point of discovery, potentially reducing the need for a customer to navigate several pages before receiving useful guidance. That creates a strong case for accurate, specific answers, but it also creates a new place where a brand's promise can be misunderstood. Tone alone cannot make the service trustworthy. A friendly answer about compatibility, returns or delivery is still a problem if it contradicts the merchant's actual terms. Enterprise teams should treat the agent as a customer-facing service that needs an approved knowledge base and clear limits. Selected retailer eligibility is an important boundary to the announcement. Brands should evaluate available participation and the supported capabilities before assuming they can launch the same experience or control every aspect of its behaviour.
How your brand can benefit / be affected
Check retailer eligibility and the data the Business Agent can actually access. Prioritise common product questions and ensure stock, specifications, offer conditions and policies have authoritative sources.
Test responses against approved product guidance and escalate material errors to the teams that own the underlying data. Track qualified handoffs and customer outcomes where available. Treat announced future capabilities as roadmap items until they are supported in your account and market.
Prioritise the questions that genuinely affect buying confidence. Document approved answers for product fit, configuration differences, delivery restrictions and common policy concerns. Identify where those answers live and who is responsible for keeping them current. If the catalogue contains several similar models, make their distinctions explicit rather than asking a conversational interface to infer them from marketing language. Include cases where the right answer is that a product is unsuitable. A useful brand agent should help the customer make a good decision, even when that decision is to choose another option or seek more information before buying.
Review the available setup against the standard you would apply to a sales or support channel. Check how information updates are handled, what can be measured and where the user goes when the agent cannot resolve a question. Test representative conversations against approved facts and note unsupported assurances. Measure qualified next actions and expectation-related support contacts alongside usage. Avoid equating a brand voice with full control of the platform experience. If access is unavailable, improving the underlying product and policy information is still worthwhile. It prepares the brand for future supported participation while reducing confusion in current search, shopping and website journeys.
News date: 11 January 2026. Editorial review: 16 September 2026. Analysis includes subsequent developments where stated.