Agentic AI is quickly becoming a core focus for financial crime and compliance teams, with new research showing growing confidence in its ability to deliver meaningful change.
Data from a joint report by Hawk and Chartis indicates that more than three quarters of compliance and risk leaders expect agentic AI to have a positive impact on financial crime and compliance effectiveness over the next two to three years.
A third believe it will lead to a major shift in how compliance work is carried out, rather than simply refining existing processes.
The findings come from AI in Financial Crime and Compliance: Charting the Path from Pilot to Maturity, which examines how banks and FinTechs are moving beyond experimentation and embedding AI more deeply into their operating models.
Optimism around the technology is translating into investment. Most compliance leaders expect spending on agentic AI to rise over the coming years, typically by up to 25%, with a smaller group forecasting even larger increases. Very few anticipate any reduction in budgets.
The anticipated impact on the financial crime workforce is more nuanced. Rather than wholesale job losses, leaders largely expect a reshaping of roles and responsibilities. Many predict moderate reductions in headcount as repetitive, manual tasks are automated, while core expertise remains essential. Others expect staffing levels to stay broadly the same, with AI changing how work is done rather than how many people are employed. A smaller group anticipates growth in specialist roles focused on oversight, governance and strategic deployment of AI.
Rather than triggering widespread job losses, agentic AI is expected to reshape how teams work. Case investigations, narrative and SAR drafting, and background research are viewed as the areas where AI agents can deliver the greatest value.
For more insights into the growth of AI, download the reports here. There are two editions to the report: Banking and Payment & FinTech.
Read the daily FinTech news
Copyright © 2026 FinTech Global
