Indosat launches AI semantic search with Google Cloud support
Indosat has launched an AI-powered Semantic Search capability in its myIM3 and bima+ applications, developed in partnership with Google Cloud and enabling greater personalisation for users seeking services and product recommendations.
This new search functionality is powered by Google Cloud's Vertex AI platform, which analyses customer behaviour and preferences. Through this technology, user searches for offerings such as data packages, lifestyle features and loyalty programmes - including IMPoin and BonsTri - can be dynamically adapted to better suit individual needs.
Search adaptation
The integration of Semantic Search allows the myIM3 and bima+ applications to interpret the semantic meaning behind user queries, enhancing the relevance and efficiency of search results. Indosat describes this as a step toward delivering more intuitive digital experiences for its customers.
Bilal Kazmi, Director & Chief Commercial Officer at Indosat Ooredoo Hutchison, said:
"With the support of Google Cloud, we are delivering a search system that is not only intelligent but also delivers tangible value, enhancing both user convenience and business performances. This collaboration represents a strategic step towards providing smarter, more personalized, and relevant digital experiences for our customers. At Indosat, we are committed to ensuring that technology makes life easier, brings services closer, and serves as a powerful tool to empower Indonesia in navigating the digital era."
The Semantic Search feature leverages insights drawn from customer data to adjust the services and recommendations offered to each user, aiming to keep pace with shifting expectations in the digital marketplace.
Google Cloud perspective
Fanly Tanto, Country Director for Indonesia at Google Cloud, commented on the deployment of Vertex AI Search in this context.
"As an industry-leading organization in Google Cloud's Indonesia BerdAIa program, Indosat has been at the forefront of innovating with advanced, enterprise AI. Vertex AI Search's in-built understanding of queries in Bahasa and English, the semantic meaning behind queries, and spelling variations will help surface more relevant digital services recommendations that meet the diverse lifestyle needs of millions of myIM3 and bima+ users. This deployment demonstrates how the unique combination of Google Cloud's open platforms and Google-quality search can be seamlessly integrated with an enterprise's custom-built AI models and existing applications, enabling smarter and more personalized digital experiences for its customers."
The system's capability to understand natural language queries - including subtle variations in spelling, as well as in Bahasa and English - aims to serve a user base with diverse linguistic backgrounds and needs.
Personalisation and efficiency
Indosat and Google Cloud highlighted that AI-powered personalisation not only improves the relevance of the service experience for each individual but also supports operational efficiency. Indosat cited a McKinsey study indicating that 71% of consumers expect personalised interactions from brands, illustrating the shift towards this kind of experience in the digital landscape.
Gamified campaign
In conjunction with the Semantic Search rollout, Indosat has launched the '30 Days of Treasure Hunt' campaign, running until mid September of 2025. The campaign is designed to familiarise customers with the new search tool through an interactive challenge. Participants are encouraged to search for specific keywords via the myIM3 or bima+ applications after receiving clues from the official Instagram accounts of IM3 and Tri. The most active users each day will have the chance to win digital vouchers, with a total prize pool of up to IDR 5.5 million.
The campaign leverages AI-powered gamification to stimulate user engagement while demonstrating the practical capabilities of the new search feature.
Language model integration
Further supporting the Semantic Search offering is Sahabat-AI, an open-source large language model (LLM) geared specifically towards the Indonesian language. This model is built to handle various local contexts, languages and dialects, aiming to improve customer interactions not just in major urban areas but also in more remote regions of Indonesia.
Indosat states that the features collectively aim to make digital activities, product discovery, and service access more meaningful for users throughout Indonesia.
