Executive Summary
A regional water utility company in Indonesia is facing major challenges in managing and accessing information spread across different departments, systems, and even physical documents. With data stored in silos, staff often spend too much time searching for information, whether it’s customer records, maintenance histories, or financial documents. This slows down operations, increases costs, and impacts customer satisfaction. Without a more efficient system in place, the company risks continued inefficiencies, reduced service quality, and difficulty in meeting regulatory or strategic needs.
To address this challenge, the company implements a centralized knowledge management solution powered by Generative AI. The solution unifies data from multiple sources into a single platform and enhances it with natural language search capabilities. Staff can ask questions in plain language and receive fast, accurate answers drawn from across the organization’s data, improving operational efficiency, decision-making, and customer service.
By adopting this AI-driven approach, the company modernizes its information management practices and establishes a scalable foundation to support future growth and innovation.
Challenge
The company encountered significant operational challenges stemming from the fragmented storage and retrieval of critical information. Data resided across disparate databases, file systems, and physical archives, impeding efficient access. This lack of a unified information repository created operational bottlenecks across various departments.
Specifically:
- Customer service representatives faced difficulties in obtaining comprehensive customer profiles, leading to extended call handling times and reduced customer satisfaction.
- Engineering personnel experienced delays in locating specific maintenance records for infrastructure assets, potentially prolonging repair cycles and increasing the risk of service disruptions.
- The finance department encountered complexities in reconciling financial transactions due to the absence of a consolidated view of relevant data.
Failure to address these challenges posed several potential negative consequences for the company:
- Increased operational expenditures due to inefficiencies in information retrieval and resource allocation.
- Diminished customer trust and satisfaction resulting from delayed responses and issue resolution.
- Impaired decision-making and strategic planning capabilities due to the lack of timely and accurate information.
- Difficulties in generating comprehensive reports for regulatory compliance and stakeholder communication.
Solution
PT. Mastersystem Infotama, Tbk. developed and implemented a GenAI-powered solution to optimize information access and management for the company. The “PT. Mastersystem Infotama, Tbk. GenAI for the company.” platform provides a centralized and intelligent interface enabling users to retrieve information from diverse data sources through natural language queries. This eliminates the need for complex database queries or manual document review, significantly streamlining information retrieval processes.
The solution architecture comprises the following key components:
- Gen AI Back End: Hosts Large Language Models (LLMs) that provide natural language processing capabilities for accurate query interpretation.
- Gen AI Front End: An intuitive user interface facilitating seamless interaction with the GenAI system for information retrieval.
- In-Memory Store: A high-speed cache for frequently accessed data, minimizing latency and improving response times.
- LLM Router: Intelligently directs user queries to the most appropriate LLM component for optimal processing.
- LLM Write Query: Translates natural language queries into SQL statements for efficient database access, eliminating the need for specialized user knowledge.
- LLM Conversational: Enables natural and interactive dialogues with the system, supporting follow-up questions and detailed explanations.
- Langchain Execute Query: Executes generated SQL queries against the database and retrieves relevant data for the user.
The PT. Mastersystem Infotama, Tbk. GenAI solution integrates seamlessly with the company’s existing data infrastructure, encompassing databases, document repositories, and other relevant systems. This integration establishes a unified information ecosystem, providing users with a single point of access to all pertinent data.
Outcome
The implementation of the GenAI-powered WhatsApp virtual assistant has significantly improved the accessibility and efficiency of billing and payment-related customer services at PDAM Tirta Raharja.
Previously, customers were required to check their billing information through the company’s website or by contacting the customer service team directly. However, many customers were unfamiliar with navigating the website, and inquiries handled by customer service agents often resulted in longer response times due to limited service hours and high volumes of requests.
With the introduction of a WhatsApp-based virtual assistant powered by Generative AI, customers can now access billing information through a platform that is widely used and familiar to most customers. Through a simple conversational interaction on WhatsApp, customers can instantly retrieve information such as current bills, payment status, due dates, and payment history.
The solution integrates directly with the company’s customer database, allowing the chatbot to deliver accurate and real-time responses. As a result, customers no longer need to navigate web portals or wait for manual responses from customer service agents.
Following the implementation, the company observed several operational improvements:
- 60–80% reduction in average response time for billing and payment inquiries compared to manual customer service handling.
- 50–70% reduction in routine inquiry workload handled by customer service agents, allowing staff to focus on more complex service requests.
- 24/7 service availability, compared to previous customer service hours that were limited to working hours.
- Higher service accessibility, as WhatsApp is already used by the majority of customers, eliminating the need to access separate web portals.
From a business perspective, the solution has improved customer service productivity, enhanced service consistency, and increased customer convenience. At the same time, it provides PDAM Tirta Raharja with a scalable digital service platform that can support additional AI-driven customer engagement capabilities in the future.