AI-generated representative image
India’s Comptroller and Auditor General is preparing to take audit-specific artificial intelligence projects nationwide, according to an official statement released on 17 November. CAG K Sanjay Murthy said the projects were ready for rollout as the department expands its use of financial data to guide audit work.
The CAG’s Audit Diwas press release records the remarks at the conclusion of the two-day observance in New Delhi. It describes readiness for deployment, rather than confirming that every audit office has already begun using the projects.
Data analysis before fieldwork
Murthy said access to digital systems lets teams examine complete transaction sets. Analysing those records before field visits helps direct enquiries and has shortened the audit cycle, the statement says. The department is building tools for auditing instead of depending solely on general-purpose AI.
A message from Prime Minister Narendra Modi, reproduced in the release, said technology can help examine large financial datasets and spot unusual patterns. It also emphasised that professional experience and independent judgement remain central to the value of an audit.
Part of a wider technology programme
The institution’s 2025–26 Impact Report roadmap places AI alongside machine learning, transaction analytics and stronger information-technology audit capability. Its proposed approach is to use these tools to identify risks, anomalies and weaknesses in controls.
That roadmap also identifies staff skills in analytics, AI and cybersecurity as areas for further investment. It discusses greater use of data from public-finance platforms and more consistent audit methods across offices. These are institutional priorities, not published performance benchmarks for the newly announced AI projects.
The distinction matters for public accountability: finding an unusual transaction pattern can guide investigation, while an audit conclusion still requires evidence and professional assessment. The 17 November announcement gives no project-by-project deployment timetable or measured accuracy results.
Read more in our Technical news coverage.

