London, UK, 12th March 2026 — Most finance teams that consider themselves AI leaders still lack the operational foundations needed to scale AI safely into core finance workflows, according to new research from Payhawk. In the study, “AI leaders” are organisations that rated their AI maturity 7–10 out of 10 (n=405 of 1,520).
The global survey of 1,520 finance and business leaders finds that while experimentation with AI is widespread, three-quarters (74%) of self-identified AI leaders do not have the governance and data infrastructure to move from adoption to operational deployment.
For finance functions, this distinction matters; AI is only meaningful when it is embedded in high-accountability processes like close, controls, approvals, exception handling, audit trails, and spend governance.
What does it take for AI to scale inside finance workflows?
Across finance organisations, AI can be piloted with minimal infrastructure. Scale, however, requires an operating stack. The research identifies five operational requirements that determine whether AI can move from ‘adopted’ to ‘operational’ inside finance workflows:
- Execution measures in place
- Minimum rules for AI use
- Skills and tools
- Budget committed
- Data usable for AI analytics.
Analysis focused on organisations that rated themselves high in AI maturity. Even within the self-identified leader group, readiness is uneven. Only 26% of AI leaders have all five requirements in place at the same time. In other words, 74% are missing at least one foundation needed to scale AI safely into core workflows.
Among AI leaders, 78% report strong availability of AI skills and tools, 69% have committed AI budgets, and 64% have execution measures in place.
The data suggests that AI adoption in finance is not being held back by a skills gap, but rather a governance and infrastructure gap. A third (32%) of AI leaders have the skills but lack minimum rules for safer use. Another 22% have implemented AI measures but still lack minimum rules to scale consistently. Meanwhile, two in five (39%) do not strongly agree that their data can support AI-driven analytics effectively.
Put simply, many teams are accumulating “rules debt” (execution outrunning minimum governance) and “data debt” (activity without reliable data foundations).
Figure 1 shows the share of AI leaders who strongly agree each scaling requirement is in place.
Full-stack readiness exists, but it is not the default
Full-stack readiness exists in the market today. However, it is a minority posture even among the most advanced organisations.
Many organisations have AI investment and governance intent, but scaling stalls ultimately due to a lack of clear minimum rules or because systems cannot reliably reconcile outputs with trusted financial data.
“In finance, AI only matters when you can delegate real work inside controlled workflows like approvals, reporting and audit trails,” said Hristo Borisov, CEO and Co-Founder of Payhawk. “Our data shows the skills and experimentation are already there. What’s missing is the operating stack, minimum rules and usable data that make AI accountable at scale.”
