A new global study by Payhawk reveals a growing gap between AI experimentation and real-world deployment in finance. Despite high adoption levels, 74% of self-identified “AI leaders” lack the governance and data infrastructure required to integrate AI into critical financial workflows.
The five pillars of AI scaling
The research outlines a clear “operating stack” required to move AI from experimentation to production use. For AI to function in high-accountability processes such as audit trails, spend governance, and exception handling, organizations must meet five key conditions:
- Execution measures already in place
- Defined rules for safe AI usage
- Access to AI skills and tools
- Committed budgets for AI initiatives
- High-quality, usable data
Currently, only 26% of AI leaders meet all five requirements, leaving most organizations unable to scale AI safely.
Governance and data vs. skills
The findings suggest that the main bottleneck is not talent, but infrastructure and governance. While 78% of organizations report having sufficient skills and tools, other areas lag behind:
- 39% lack confidence in their data readiness for AI analytics
- 32% have skills but lack clear usage rules
- 22% have execution in place but lack scaling frameworks
This imbalance creates what the report describes as “rules debt” and “data debt,” where AI adoption outpaces governance and reliable data foundations.
Insights from leadership
Hristo Borisov, CEO and Co-Founder of Payhawk, noted that AI only delivers value in finance when it can be trusted with real operational tasks within controlled environments.
The study, conducted with IResearch, surveyed 1,520 senior professionals across eight countries, highlighting the operational challenges CFOs face as AI adoption accelerates in 2026.
