The Canada Revenue Agency (CRA) is turning to artificial intelligence and enhanced training to address widespread issues with the accuracy of responses provided by its call centre agents.
Assistant Commissioner Melanie Serjak told MPs this week that the CRA plans to roll out a “senior and standardized” training model and integrate AI-based tools to help agents deliver more reliable tax information to Canadians.
The move follows a scathing Auditor General’s report revealing that CRA agents correctly answered only 17 per cent of tax-related questions, while the agency’s chatbot “Charlie” managed 33 per cent accuracy.
Critics have said the CRA has prioritized adherence to employee schedules over service quality.
In response, Serjak said the CRA is “mobilizing the entire agency” to meet its 100-day improvement plan set by Finance Minister François-Philippe Champagne, with progress already underway.
The CRA maintains that automation and quality review improvements will help restore public trust, service reliability, and transparency in taxpayer support.