Accenture Song, Accenture’s customer experience division, outlined company strategies for agentic commerce at a media briefing on September 2, 2026. Senior Manager Tatsuro Ishii said people and AI agents will both take part in purchasing decisions, so products need to be selected by both.

Ishii said AI assistants now search for products in response to questions, but may evolve in stages into agents that recommend products based on accumulated information and then into autonomous agents that handle payment. Systems for completing purchases within conversational AI services such as ChatGPT, along with standards connecting AI agents to retailers and payment providers, are being developed. Ishii said AI services are becoming a key battleground for customer purchases and urged companies to prepare early.

The first strategy is to make product information easier for AI agents to read. Alongside AIO (AI optimization), which standardizes product names, prices and specifications, companies should provide information about intended users and use cases. For beer, that could include target customers, taste, occasions for drinking and food pairings, in addition to alcohol content and volume. Ishii said information identifying a product as suitable after surfing or as low-calorie could help an AI agent select it for those situations. Publishing information on a company website alone is not enough: companies also need to submit it through product registration channels run by AI service providers and support common data-sharing standards. Accenture Song said companies need a system to gather information from relevant parties, understand the standards and specifications, and keep expanding the information over time.

The second strategy is to build a brand-owned agent using the company’s knowledge to help customers find products, get post-purchase support and consult with stores. Accenture Song said companies should analyze conversation data after launch to improve the accuracy of responses and recommendations. Its example, JINS AI from eyewear chain JINS, is a customer-service agent trained on a database of store staff conversations, including knowledge about combinations of lenses and frames. It can answer questions about products and whether items can be combined. During busy periods, customers can use it to organize what they want before speaking with store staff.

Ishii said which strategy matters more depends on the industry and product. For fast-changing goods such as daily necessities, companies need to update information quickly so AI agents can select their products. For high-priced goods or services that require specialized explanations, he said, providing suitable information through a brand-owned agent can help differentiate the business.

Accenture Song, Accenture’s customer experience division, outlined company strategies for agentic commerce at a media briefing on September 2, 2026. Senior Manager Tatsuro Ishii said people and AI agents will both take part in purchasing decisions, so products need to be selected by both.

Ishii said AI assistants now search for products in response to questions, but may evolve in stages into agents that recommend products based on accumulated information and then into autonomous agents that handle payment. Systems for completing purchases within conversational AI services such as ChatGPT, along with standards connecting AI agents to retailers and payment providers, are being developed. Ishii said AI services are becoming a key battleground for customer purchases and urged companies to prepare early.

The first strategy is to make product information easier for AI agents to read. Alongside AIO (AI optimization), which standardizes product names, prices and specifications, companies should provide information about intended users and use cases. For beer, that could include target customers, taste, occasions for drinking and food pairings, in addition to alcohol content and volume. Ishii said information identifying a product as suitable after surfing or as low-calorie could help an AI agent select it for those situations. Publishing information on a company website alone is not enough: companies also need to submit it through product registration channels run by AI service providers and support common data-sharing standards. Accenture Song said companies need a system to gather information from relevant parties, understand the standards and specifications, and keep expanding the information over time.

The second strategy is to build a brand-owned agent using the company’s knowledge to help customers find products, get post-purchase support and consult with stores. Accenture Song said companies should analyze conversation data after launch to improve the accuracy of responses and recommendations. Its example, JINS AI from eyewear chain JINS, is a customer-service agent trained on a database of store staff conversations, including knowledge about combinations of lenses and frames. It can answer questions about products and whether items can be combined. During busy periods, customers can use it to organize what they want before speaking with store staff.

Ishii said which strategy matters more depends on the industry and product. For fast-changing goods such as daily necessities, companies need to update information quickly so AI agents can select their products. For high-priced goods or services that require specialized explanations, he said, providing suitable information through a brand-owned agent can help differentiate the business.