DEWA opens AI gateway for public services

Dubai Electricity and Water Authority has rolled out direct access to its customer services through ChatGPT, positioning itself as the first government entity and the first utility globally to integrate end-to-end service delivery within a generative AI platform. The move allows customers to obtain information, complete routine requests and navigate account services through natural language prompts, bypassing traditional websites and call centres.

The integration, launched as part of Dubai’s wider digital government push, enables users to interact with DEWA’s official AI assistant inside ChatGPT. Customers can ask about electricity and water bills, consumption patterns, connection procedures, service activation timelines, conservation programmes and sustainability initiatives, with responses drawn from DEWA’s verified data environment. The authority says the system has been designed to provide accurate, policy-aligned answers while maintaining data privacy and security standards required of a critical utility.

Officials involved in the project describe it as a shift from information publishing to conversational service delivery. Rather than searching webpages or mobile app menus, customers can frame queries in everyday language and receive guided responses that mirror a live service interaction. The AI assistant is also structured to redirect users to secure transaction channels when authentication or payments are required, ensuring sensitive operations remain within controlled systems.

DEWA’s adoption of generative AI follows sustained investment in digital infrastructure, automation and predictive analytics across its operations. The utility has already deployed AI and machine learning in grid management, demand forecasting, asset maintenance and customer experience tools. Embedding services into ChatGPT extends that strategy to public-facing engagement, reflecting a belief that conversational interfaces will become a primary access point for digital services.

The initiative aligns with Dubai’s ambition to be a global hub for artificial intelligence and smart government. The emirate has encouraged public entities to experiment with AI to improve efficiency, reduce service friction and enhance accessibility. By moving services directly into an AI platform with a large global user base, DEWA is effectively meeting customers where they already spend time, rather than requiring them to learn new interfaces.

Industry analysts view the move as significant for utilities worldwide, a sector often criticised for slow digital adoption. Utilities typically rely on legacy systems and conservative service models due to regulatory oversight and infrastructure complexity. A successful AI-driven interface at the scale of Dubai’s power and water network could provide a template for peers exploring similar deployments, particularly in cities pursuing smart infrastructure agendas.

At the same time, experts caution that AI-mediated public services carry governance risks. Ensuring responses remain accurate, up to date and free from unintended bias is critical, especially when customers rely on them for billing, compliance or service obligations. DEWA has indicated that the AI assistant is trained and monitored using approved content, with continuous oversight to prevent misinformation and to align outputs with official policies and tariffs.

Privacy and cybersecurity are also central considerations. Utilities manage sensitive personal and consumption data, and any AI integration must safeguard that information against leakage or misuse. DEWA says the ChatGPT interface does not expose confidential data and that any account-specific actions require secure authentication through official channels, separating conversational guidance from transactional systems.

The launch comes as governments globally test generative AI for citizen services, from chatbots handling municipal queries to AI-assisted licensing platforms. What distinguishes DEWA’s approach is the depth of integration and the decision to operate directly within a widely used AI model rather than a standalone chatbot. This approach reduces friction for users but increases the importance of rigorous controls and transparency.