Roughly 95 per cent of organisations see no measurable return on their generative AI investments, but finance functions are achieving substantial value by embedding AI into automated workflows rather than running experimental pilots, according to research from MIT and Anthropic.
MIT’s State of AI in Business 2025 shows the earliest and clearest returns from AI are appearing in support functions such as finance, procurement and operations through reduced external spending, faster cycles and tighter controls, reports Bain & Company.
Michael Heric, a Bain & Company partner, said the MIT analysis shows the problem is not AI models but the approach. Companies that reap the biggest benefits embed AI into real workflows, ensure systems learn from feedback and measure business outcomes.
Heric said the report’s message for CFOs is to “spend differently” rather than spend less on generative AI. He advised finance leaders to aim more investment at use cases that affect cash, cost and risk, and to focus on business outcomes rather than what AI tools do.
Finance processes tend to be repeatable, data-rich and policy-bound, the conditions in which generative AI delivers substantial benefits. However, budgets remain skewed toward visible, top-line pilots while back-office automation, where the payback is often faster, remains underfunded.
Enterprise leads the way
Anthropic’s Economic Index reveals that 77 per cent of its enterprise API usage is linked to automating tasks, with that share rising over time as users shift from iterative prompting to clear, directive commands. The behavioural shift suggests the most effective use of enterprise AI is automation, not chat-related assistance.
Anthropic’s research found the main limitation is not price but data that provides context. When companies give AI complex tasks, they tend to provide much longer inputs, but every one per cent increase in input length yields only about 0.38 per cent more output. Demand does not change significantly when token prices change, suggesting cost is not holding back adoption.
In finance, the main challenge is building a solid data foundation, including well-managed data products for the chart of accounts, vendor and customer records and policy documents. These need to connect with enterprise resource planning and enterprise performance management systems while ensuring data is retrieved securely and within proper boundaries to remain compliant with the Sarbanes-Oxley Act and International Financial Reporting Standards.
Leading companies are moving from application-centric systems to data-driven platforms, then to intelligent orchestration, and ultimately to AI agents that manage multistep workflows. The goal is a finance function that operates continuously, where closing, forecasting and review processes run in real time and human attention is needed only for exceptions.
