AI is everywhere at work. Except in company finance. Why?

In MIT’s hot report ‘State of AI in Business 2025’, workers from over 90% of the companies surveyed reported regular use of personal AI tools for work tasks.

Generative AI has rewired knowledge work:

• Pinstripe lawyers use ChatGPT all day.
• Emergency room doctors grab fast second opinions.
• Management consultants whip up board decks.
• Jedi software developers ship code with copilots.

What’s the equivalent behavior for CFO/finance teams at companies? What is widely adopted?

Nothing.

Really? Why isn’t finance keeping pace?

A few reasons.

• Different medium: GenAI Large Language Models (LLMs) are fluent in language; finance runs on ‘structured data’: hundreds of disjointed tables across enterprise systems of record, filled with rows and columns of numbers.
• Privacy and access: LLMs are smart: they’ve read every book at the library, every study online; they have vast general business knowledge; and since 2023, they reason, a bit like humans. But… models have never seen YOUR internal data! And most enterprises will never share their data to train a model.
• Trust requirements: Finance needs transparent logic and math and Excel-style receipts, not black-box answers or cryptic database code.

These challenges apply to business and operations teams that also rely on key enterprise systems.

What do leaders actually want?

‘Can I ask core systems plain-English questions and get answers grounded in source data, with traceable logic and exportable support?’

Disruption is inevitable in enterprise finance.

The question is not if; it’s how.

Maxa bridges the gap between language models and enterprise finance systems. See how CFOs are asking plain-English questions and getting traceable, exportable answers at maxa.ai.