Back

How the Department of Statistics Malaysia turned days-long public data inquiries into instant answers with Chatbase

Department of Statistics Malaysia logo dark

Days to seconds

Response time for public inquiries

Near zero

Manual inquiries now reaching staff

2

Public dashboards served by AI agents

At a glance

The Department of Statistics Malaysia (DOSM) is the national government agency responsible for collecting, interpreting, and disseminating the country's official statistics. In recent years it has become one of the most digitally advanced statistics offices in the world: in the 2024/25 Open Data Inventory, Malaysia ranked first globally, up from 67th two years earlier. To make the information behind its public dashboards easy for the public to reach, they deployed two AI agents built on Chatbase, AgroBot and PetroBot, as a 24/7 first line of assistance.

Rather than replacing human support, the chatbot serves as a first line of assistance by handling common inquiries 24/7, improving accessibility to official statistical information while reducing repetitive questions for our team.

Michael Immanuel, Assistant Director, National Big Data Analytics Centre, DOSM

Making official statistics accessible to everyone

Before Chatbase, the public had no direct, convenient way to ask questions about DOSM's statistics and services. People searched across multiple webpages and documents, or contacted the department by email and phone. Responses took around one to two working days and depended entirely on staff availability. Now the agents answer in seconds, day or night, and manual email inquiries about these dashboards have dropped to almost nothing. That gives the public instant access to official statistics and gives DOSM staff their time back for more complex, higher-value work. For an agency measured on access rather than revenue, that is the return.

Before Chatbase, the public did not have a direct and convenient channel to ask questions about statistical data or DOSM services. Most inquiries relied on traditional communication methods, which could be slower.

Michael Immanuel, Assistant Director, National Big Data Analytics Centre, DOSM

Two purpose-built agents

DOSM evaluated several outsourced chatbot solutions, but found the responses inconsistent and the customization limited. Chatbase let the team build on its own knowledge base with no heavy development. Because DOSM runs two very different public dashboards, they built a dedicated agent for each, trained only on that dashboard's data, so answers stay accurate and specific rather than one bot guessing across everything. Both went live within a day, answering in English and Malay.

  • AgroBot, on the TaniStats dashboard. TaniStats is DOSM's agricultural statistics gateway, pulling farm, food, and supply-chain data into one place.
  • PetroBot, on the Pantau Krisis dashboard, which tracks the impact of the global supply crisis on Malaysia, things like fuel prices, cost of living, and currency movements, in near real time.

It was significantly easier to customize, integrate with our existing knowledge base, and maintain. Most importantly, it delivered faster, more accurate, and more natural responses, providing the public with a much better experience when accessing DOSM information.

Michael Immanuel, Assistant Director, National Big Data Analytics Centre, DOSM

What's next

DOSM currently runs the agents as standalone assistants on its dashboards. Next, the team is looking at connecting Chatbase to other DOSM platforms and internal systems, so the agents can do more than answer questions and start linking the public directly to the services behind the data.

As our implementation evolves, we are exploring opportunities to integrate Chatbase with other platforms and internal systems to further enhance the user experience and streamline support workflows.

Michael Immanuel, Assistant Director, National Big Data Analytics Centre, DOSM

Better support starts today

Join thousands of teams using Chatbase to build AI agents that deliver exceptional support at scale — across chat, email, and voice.