CAVIA Phase 2 is an intelligent search and analysis system developed by the Canada Revenue Agency to help employees more efficiently access and utilize diary notes stored in the ACSES system. The system uses advanced natural language processing to search across all diary notes and extract relevant information about previous account actions, such as payment requirements issued, phone calls made, associated account numbers, and legal warnings.
This system is currently in development and will be used by Canada Revenue Agency employees working in collections and verification. The system processes personal information from individual taxpayers, corporations, trusts, and other tax entities. Employees will be able to search diary notes more quickly and accurately, improving productivity while reducing manual research time.
Key transparency information: This system involves processing personal information, uses natural language processing and generative AI capabilities, and is being developed by a vendor (AWS). The Canada Revenue Agency will notify the public about this AI system in accordance with government transparency requirements.
