Use of AI to support climate services at MET Norway

User enquiries are an important part of operational climate services at The Norwegian Meteorological Institute (MET Norway). Requests are received through email and telephone, and often concern past weather and climate conditions in connection with media requests, insurance claims, or general documentation needs. Many of these enquiries require manual retrieval of data, calculation of statistics, and identification of relevant datasets, webpages, or standard response formulations, making the work both time-consuming and resource-intensive.

To support this work, MET Norway is exploring the use of artificial intelligence (AI) in the form of a chatbot designed to assist with routine enquiries. In the first phase, the focus has been on insurance-related cases, where users need documentation of past weather events associated with damage to property, for example wind damage. 

The chatbot is being developed by a team of students, while domain experts at MET Norway contribute by defining operational needs, identifying relevant use cases, and testing prototypes throughout the development process. This collaboration helps ensure that the tool is grounded in real workflows and addresses practical needs within climate services.

The aim of the project is to assess whether an AI-based assistant can improve efficiency and support service quality by helping staff retrieve relevant information, guide users to the right data sources, and draft responses to recurring enquiries. At the same time, the project provides an opportunity to identify important limitations and clarify where human expertise remains essential.

The study represents a practical case of how AI can be introduced into operational climate services. If the approach proves effective, the service could later be extended to other types of enquiries and, in the longer term, potentially developed into an external service for users.