100% of firms expect positive outcomes from agentic/predictive AI, but adoption rates lag
Written by Iain Armstrong
- Insights
- Featured Insights
Excited conversations and ambitious statements about the transformative potential of AI are nothing new in compliance.
This year, however, things are different. AI adoption has become the baseline for most compliance teams. In our latest global survey of 600 compliance decision-makers, 93% reported currently using, piloting, or evaluating a ‘standard’ AI solution for customer screening, and a similar number, 87%, are using or evaluating AI solutions for transaction monitoring.
In other words, AI has gone from a range of technologies with untapped potential to an essential feature of efficient, growth-friendly compliance programs. However, this also means that in an increasingly competitive financial services landscape, basic AI adoption – involving automation in silos, unintegrated workflows, or superficial AI overlays on still-fragmented architecture – is no longer enough.
Now that AI in compliance has become the new normal, forward-thinking firms should be proactively assessing how the most advanced AI capabilities can further enhance their compliance programs. Agentic AI, referring to AI architecture with coordinated autonomous systems capable of automating entire workflows without direct human supervision, is foremost among these, with Level 1 alert remediation a clear, high-value use case.
Glossary
Agentic AI describes AI models that are equipped with agency: the ability to plan, reason, and execute tasks in pursuit of objectives. Unlike ‘standard AI’ systems that are limited to a single output (eg. a model that flags a transaction as suspicious based on a pre-set rule), agentic AI can interact with external tools, adapt behavior based on context, and operate iteratively until the desired outcome is achieved. An example of agentic AI is the auto-remediation of level 1 alerts – where the AI not only identifies a low-risk alert but also independently gathers the necessary data, drafts the closure narrative, and closes the case without human intervention.
Predictive AI forecasts future outcomes by analyzing data patterns, while agentic AI acts autonomously to achieve goals by combining predictive capabilities with autonomous planning, decision-making, and action-taking in an environment, often involving multiple agents working together with human guidance or oversight. The key difference is prediction (what might happen) versus autonomous action (making it happen).
According to our survey, 100% of organizations have achieved or expect to see positive outcomes from adopting agentic/predictive AI. However, this is not yet reflected in adoption rates, with only 33% currently using an agentic solution for customer screening and 32% for transaction monitoring.
The results are unequivocal: AI use is now a table-stakes requirement. For the minority of organizations still on the sidelines, the message is clear – if you aren’t using AI for core functions like screening and monitoring, you are already operating at a competitive and efficiency disadvantage. With early adopters best positioned to make significant efficiency gains from agentic AI, it will be interesting to see how these numbers look in a year’s time.
The State of Financial Crime 2026
Get insights on financial crime trends from our global survey of 600 senior decision-makers and expert guidance from our Financial Crime Compliance Strategy team. Coming February 2026.
Pre-register nowOriginally published 17 December 2025, updated 13 January 2026
Disclaimer: This is for general information only. The information presented does not constitute legal advice. ComplyAdvantage accepts no responsibility for any information contained herein and disclaims and excludes any liability in respect of the contents or for action taken based on this information.
Copyright © 2026 IVXS UK Limited (trading as ComplyAdvantage).