Gill Capital Group, a retail brand development and distribution company, has piloted a generative AI-powered search agent and a virtual shopping assistant on H&M’s e-commerce platforms in Indonesia and Thailand.
The company has piloted the tools with a test group of shoppers.
Early trials have shown strong results in boosting engagement and sales.
The company now plans a full rollout for Indonesia and Thailand later this year, before extending the use of AI agents to other brands in its portfolio.
According to the company, the search agent helps fix one of retail’s biggest challenges, which is poor search functions that fail to understand what shoppers really want.
This AI agent understands the true intent behind search queries in natural language.
It understands what shoppers mean, even if they search in local languages, English or make a typo - like asking for a blouse for hot weather in Thai or an elegant tunic for Eid in Bahasa.
By focusing on the meaning of words, not just keywords, it surfaces contextually relevant results.
On the backend, the agent also automatically analyses and organises a vast product catalogue.
This frees staff from the task of researching and adding keywords to each product.
Gill Capital is working with Google Cloud, and the new AI search agent is built on Vertex AI Search for Commerce.
Integrating a conversational shopping agent
Gill Capital is also integrating a conversational shopping assistant into H&M’s websites and apps.
Developed using Vertex AI Search for Commerce, the agent offers personalised recommendations and can handle complex queries like checking if an item is in stock at nearby stores.
It also answers customer service questions, including return policies, creating a seamless bridge between digital and physical shopping.
Gill Capital Group’s group chief data and analytics officer, Victor Siow, said, “While search technology has advanced from keyword-matching to semantic and agentic search, many online retailers haven’t kept pace. With Google Cloud, we’re utilising the same underlying components that power Google Search today to gain a competitive edge.”
By grounding reasoning models like Gemini 2.5 Flash on company data sources, the company assures that AI agents provide relevant and accurate outputs, Siow said.
As part of the AI strategy, the company is bringing horizontal search and conversation use cases to other brands while pursuing new applications in key areas like supply chain optimisation, Siow added.
The solutions are being piloted through Google Cloud’s AI Cloud Takeoff (AI CTO) program, launched with Digital Industry Singapore (DISG).
