Contact a consultant

ai

Hakuhodo and Shopify launch Data Enrichment so AI agents can understand products

Hakuhodo and Hakuhodo DY ONE said in Tokyo on 16 November that they had launched Data Enrichment on Shopify, a Shopify-integrated tool developed with Shopify Japan for agentic commerce.

Promotional artwork for Hakuhodo's announcement of Data Enrichment on Shopify.
Image: Hakuhodo

Key takeaways

  • Hakuhodo and Hakuhodo DY ONE launched Data Enrichment on Shopify with Shopify Japan on 16 November.
  • The service rewrites product specifications into situational value that AI agents can interpret; the core technology is patent pending.
  • Enriched data is written into Shopify product masters, descriptions and metafields, with a link through to Shopify checkout.
  • Hakuhodo illustrated the idea with a washing machine's 32 dB noise level and heat-pump drying, mapped to quiet overnight laundry for dual-income households.

Why it matters

When shoppers describe a real-life need to an AI agent, they are more likely to be shown products that fit the situation, not just models with similar specifications. For brands and retailers, product data that lists only technical attributes may not give an agent enough context, so a suitable item can be left out of the recommendation list.

What the service does

Tokyo-based marketing company Hakuhodo, together with Hakuhodo DY ONE, announced on 16 November the launch of Data Enrichment on Shopify. The service was built on Shopify's commerce platform in collaboration with Shopify Japan K.K., and sits inside the Hakuhodo DY Group's Agentic Commerce ONE suite.

Hakuhodo said generative AI and AI agents are giving rise to a way of shopping in which people express an intention or a problem, and an agent searches, recommends and even buys on their behalf. Unlike keyword search, agents interpret context and purpose. If a company only provides functional specifications and technical attributes, a product that would actually suit the user may still be excluded from the recommendation list.

How specifications become situations

At the core of the service is a patent-pending technology. Hakuhodo said it extracts the experiential value that people seek and automatically generates a structured "logic bridge" between product specifications and that value.

In the company's example, a washing machine's "32 dB operating noise" and "heat pump drying system" can be connected to the idea that a dual-income household can finish washing and drying quietly at night, regardless of the weather. That structured information is stored in the product database and incorporated into the product master, so an agent can interpret the user's situation more accurately when someone asks for help.

The Shopify connection

Enriched product information is reflected in and centrally managed within Shopify, including product master data, descriptions and metafields. Hakuhodo said companies can optimise product-detail-page data for external AI environments, increasing discoverability and creating a path from an AI recommendation to Shopify's purchase experience without extensive extra development.

The service is also designed to validate recommendation accuracy with test queries, monitor traffic contributed by external AI platforms, and keep tuning product data based on those results. After an agent recommends a product, a link can take the shopper into checkout using Shopify's payment infrastructure.

Hakuhodo also noted that Shopify has co-developed and natively adopted the Universal Commerce Protocol (UCP). As support for external AI platforms expands in Japan, that foundation is intended to help with capabilities such as real-time inventory checks and AI-completed purchases. The announcement did not disclose pricing, launch customers or a rollout timetable.

Background

Agentic commerce describes a pattern in which a person tells an AI agent what they need, and the agent searches, compares and even pays. Hakuhodo said conventional product masters often emphasise functional specifications and may not give agents enough contextual information to judge whether a product fits a particular situation.

What to watch

  • How many Shopify merchants adopt the enrichment service, and whether it expands beyond Japan.
  • Whether test queries and traffic monitoring show that richer product data actually improves the chance of being recommended.
  • How Hakuhodo will handle personal data and consent if it later expands into customer-data enrichment, as it has planned.

Sources

  1. Hakuhodo

Questions

Has the technology already been granted a patent?

No. Hakuhodo described it as patent pending. The announcement did not say a patent had been granted.

Where does the shopper pay?

According to Hakuhodo, after an AI agent recommends a product, a link provided by the agent can take the shopper into checkout that uses Shopify's payment infrastructure.

News