Three-quarters of finance teams now use AI, but most lack readiness for critical audits
Seventy-five percent of finance leaders are deploying AI tools in their operations, according to new research from KPMG. The adoption rate has nearly tripled since 2024, when just 30% of finance teams reported using the technology.
Yet readiness lags far behind deployment. Only 42% of finance leaders say their AI systems are "strongly ready" for assurance work - the process of verifying financial statements and assessing a company's ability to cover future liabilities.
The accuracy problem
AI systems regularly produce inaccurate output. Research shows AI-generated content contains errors or "hallucinations" - approximations used to fill data gaps - in as many as 45% of cases.
Finance cannot absorb that error rate. In 2023, the US Financial Stability Oversight Council classified AI as an "emerging vulnerability" in the financial sector, citing risks to data privacy and the potential for hallucinations to complicate operations.
The risk became concrete when EY published a report on cybersecurity that was more than 70% AI-generated. The system invented citations to Forbes, McKinsey, Gartner, TechCrunch, and WIRED that did not exist. EY removed the report from circulation.
ROI claims outpace readiness
Despite these concerns, 71% of finance leaders report positive returns on their AI investments. That figure is higher than in most other sectors, suggesting finance teams see genuine productivity gains.
The gap between adoption and assurance readiness creates exposure. At least some portion of the 75% of firms deploying AI in finance overlap with the 58% who admit they are not adequately prepared to verify the quality of their AI-assisted financial statements.
Assurance readiness matters
Organizations that can produce AI audit evidence efficiently report dramatically better results. They achieve error reduction rates of 33% compared to 6% at firms without that capability. They also report 42% confidence in scaling AI versus 14% elsewhere.
Data quality is both the biggest barrier and the biggest opportunity. Thirty-six percent of organizations cite improving data quality, integration, and system interoperability as their greatest chance to extract more value from AI in finance - and as a major vulnerability.
For finance leaders, the message is clear: speed of deployment matters less than readiness to verify what the system produces. AI for Finance requires building governance and audit capabilities alongside adoption, not after.
Those planning AI implementation should consider the AI Learning Path for CFOs, which covers financial strategy, forecasting, and assurance-ready AI implementation.
Your membership also unlocks:
