WaterCopilot is an AI-powered virtual assistant developed to support smarter, faster, and more accessible water resource management. Built collaboratively by the International Water Management Institute (IWMI) and Microsoft for the Limpopo River Basin, and adaptable to other regions, it integrates trusted datasets, real-time APIs, and AI-driven insights into one intuitive conversational platform.
What You Can Do With WaterCopilot:
- Ask questions in plain language — "Show rainfall trends near Pafuri for the last 3 months" or "What is the current e-flow status across all monitoring sites?"
- Explore the basin at a glance — Interactive maps, charts, and visualisations covering river flows, rainfall, reservoirs, water accounting, and environmental flow indicators.
- Get intelligent alerts — Automated threshold notifications for environmental flow compliance, rainfall events, and reservoir level changes.
- Access reliable, referenced data — Every response cites its source, whether from APIs or indexed policy and technical documents, so information can be verified and traced.
- Generate summaries and visuals — Instantly produce shareable charts, maps, trend analyses, and briefs for teams and stakeholders.
- Search policy and technical documents — Semantic search across indexed Limpopo Basin policies, LIMCOM reports, and technical publications.
Who It's For: Water managers, catchment management agencies, basin authorities, researchers, NGOs, farmers, and anyone involved in water governance who needs timely, understandable water information for planning, monitoring, and policy support — regardless of their technical background. Multilingual support in English, Portuguese ensures accessibility across the basin and beyond.
Part of the Limpopo Digital Twin: WaterCopilot is one component of IWMI's Limpopo River Basin Digital Twin platform, connecting to geospatial visualisation, hydrological modelling, the Open Data Cube for satellite-derived analysis, and operational forecasting systems.
Technology Behind It: Powered by GPT-4, tool-calling architectures, and Retrieval-Augmented Generation (RAG) via Azure OpenAI Services, WaterCopilot bridges human-AI collaboration for water resource management. Two specialised plugins underpin the system: the iwmi-doc-plugin for semantic document search via Azure AI Search, and the iwmi-api-plugin for querying hydrological databases and generating visualisations on demand.
Privacy and Permissions: Location, notifications, microphone, and gallery access are optional and only used to enhance specific features such as map visualisation, alerts, voice-based input, and image analysis. WaterCopilot respects user privacy and does not collect or sell personal data. All data transmissions are encrypted in transit.
Disclaimer: WaterCopilot provides information to support water resource management and does not replace official forecasts, warnings, or directives from national or regional authorities. Always follow guidance from local hydrometeorological and disaster management agencies.
https://digitaltwins.iwmi.org/innovations/water-copilot/
Requires an active internet connection. Feature availability may vary based on data coverage across sub-basins.