Last updated: 23 Jun 2025 09:00 Posted in: AIA
Henry Wyard explains how AI can help accountants to become aware of financial crime, and the role it can play in tackling it.
In October last year, I had the pleasure of writing in International Accountant magazine about the potential that artificial intelligence (AI) has to transform the way in which organisations protect themselves, their economies and their societies from financial crime. In that article, I outlined the key capabilities of AI technology, which give it such promise for new anti-financial crime applications.
But some questions were left unanswered. How are organisations endeavouring to transition from promise to practice? How are those responsible for anti-financial crime efforts – from leaders of regulated businesses to directors of law enforcement agencies – hoping to turn the transformative potential of anti-financial crime AI into reality?
The results of our survey
To answer these questions, we at Themis, a global financial crime research and technology specialist, launched a survey of CEOs, boards and executive leadership teams (across industries, in public and private sectors) to find out how they are approaching the evolving role of AI in anti-financial crime.
Responses to this survey underpin our latest report, ‘AI in anti-financial crime: the state of adoption in 2025’. We posed nine questions to a hand-selected sample of 74 organisational decision-makers in jurisdictions across the globe, assessing the current rate of adoption of AI, uncovering specific needs in this domain, and identifying emerging trends that will likely shape the future of the use of AI in anti-financial crime practices.
Our analysis of responses revealed three key findings about the current role of AI in anti‑financial crime practices:
What do these findings mean in detail? Let’s unpack them.
A wave of AI adoption is coming
According to our respondents, AI plays a limited role within their current usage of anti-financial crime and compliance technology, with 69% of those surveyed not currently using AI for these purposes. This appears set to change rapidly in the next few years – 51% of respondents not currently using AI stated that they will acquire AI within the next three years; and over 80% of respondents plan to have acquired new AI-driven anti-financial crime systems by 2030.
The backdrop to this is the fact that legacy technology presents challenges to organisations. Existing technology systems were cited by 39% of respondents as a significant obstacle to effective anti-financial crime and compliance – but more than three quarters of these respondents do not currently use any AI.
If, therefore, AI can resolve the issues that limit the effectiveness of legacy technology systems, it will revolutionise a key problem area in anti-financial crime.
Judging by responses from the organisations that do have AI-enabled anti-financial crime systems, this potential is already being realised. Respondents who do use AI tools stated that their organisations use AI in eight separate categories of activity that cover the full range of anti-financial crime and compliance processes, including Know Your Client and due diligence checks, adverse media screening, customer risk scoring and investigations.
This all suggests that while AI may not currently be used very widely in anti-financial crime and compliance, it is being used in a rich variety of ways, with AI tools already in use for a diverse set of different processes across financial crime and compliance practices. Future implementation of AI therefore has the potential to be carried out broadly, across different areas of organisations and different compliance processes, rather than narrowly, in a limited domain of activity.
This is good news for organisations – a ‘cross‑functional’ model of AI implementation was described in the Harvard Business Review in 2019 as a key method to achieving the biggest impact from AI adoption.
Senior leaders are ready
The survey’s second key finding is that many senior leaders understand how AI can improve and innovate across anti-financial crime practices.
85% of respondents believe their leadership has at least an intermediate understanding of AI’s risks and opportunities for anti-financial crime, with 50% rating this understanding as somewhat or extremely high. Although these ratings were largely self-reported by senior leaders, and should be treated with a degree of caution, at the very least they underline that the leadership of many organisations are actively considering AI for anti‑financial crime.
But what AI solutions do senior leaders want? By a significant margin, the most common reason that respondents gave for why they would procure new AI technology was ‘to save time and cut costs’. Respondents saw AI as presenting their organisations with the opportunity to develop cheaper and better solutions to the existing challenges they face; in particular, the significant challenges they face from outdated legacy technology.
Respondents also highlighted AI’s potential to develop entirely new capabilities in financial crime investigation and analysis, with AI unlocking pathways for organisations to innovate, as well as improve, their processes. The desire for new capabilities appears to be driven by an understanding that the risk landscape is rapidly evolving; the second most common reason for procuring new AI technology was ‘to deal with new types of risk’, many of which will also be AI‑powered.
The adoption of AI for anti-financial crime purposes will, therefore, be driven by a mixture of a desire to improve existing processes and a need to develop new capabilities. Although the procurement of AI was most commonly seen by respondents as of medium priority to their organisations, the urgency of AI adoption will likely be greatly increased as the ‘new types of risk’ that participants identified as a motivating factor become more prevalent and sophisticated.
Perceived barriers to adoption can be overcome
Despite the optimism around the potential for new transformative tools, organisations still perceive significant barriers to adopting AI for anti-financial crime. Broadly, perceived barriers can be split into three categories:
The most significant concern respondents had regarding AI tools for anti-financial crime and compliance purposes concerned reliability and accuracy. Given the serious and highly sensitive nature of many anti-financial crime processes, these issues are crucial, and are being addressed within global efforts to build trustworthy AI systems. While some respondents additionally expressed doubts that the costs of new AI technology would be justified by the benefits, it must be noted that much research, such as a 2024 whitepaper published by academics at the University of Strathclyde, has found that AI can deliver ‘greater efficiency at reduced costs for organisations’ financial crime risk management.
Issues within organisations related to a lack of awareness and knowledge concerning AI for anti-financial crime. 45% of respondents cited knowledge gaps among staff as a major obstacle for their organisations procuring new AI tools; this was a far more significant issue for our research population than AI scepticism, which only 16% identified as a concern.
Regulatory uncertainty was another key concern, with over a third of respondents citing this as a barrier. Notably, this concern was shared by both those with self-professed high AI literacy and those with moderate to low understanding. This has some justification: regulatory frameworks for AI are still in their infancy, with many jurisdictions lacking comprehensive legislation to govern AI usage. Nevertheless, the principles of AI regulation have been articulated well; the UK regulatory principles outline clearly the areas to be addressed, while regulators like the Financial Conduct Authority have driven forward initiatives to propel the development and adoption of new AI tools.
As the regulatory and technology landscapes continue to develop, many of these issues will quickly be overcome. While some of the more existential concerns around AI (e.g. the trustworthiness of AI tools) may expected to remain for the long term, they should not be seen as fundamental barriers to adoption. By building familiarity with existing and developing AI applications for anti-financial crime, organisations can position themselves to take advantage of the many new opportunities they are presented with.
What should organisations do next?
So, what actions should organisations take in response to these findings? Our report, which can be accessed for free via contains 10 key recommendations to guide those looking to apply AI in their anti-financial crime practices.
But to end with an overarching recommendation, I’ll briefly set out Themis’s own approach to applying AI in anti-financial crime. We start from a ‘problem-first’ perspective, combining the knowledge of our human financial crime experts and compliance practitioners, who have lived and breathed the challenges that organisations face when confronting financial crime risk, with ground-breaking AI capabilities delivered by our development team. The blending of human and artificial intelligence allows us to build solutions that make a meaningful difference in the fight against financial crime).
The principles of our approach, which applies human expertise and advanced AI to develop solutions targeted at the core needs of organisations facing financial crime risk, are applicable to many others. This survey underlines that we stand at a critical juncture in the progression towards an AI-enabled anti‑financial crime future: now is the time to move from recognition to action, from being aware of AI’s potential to adopting it in practice.
Our efforts to reach out to business leaders were kindly supported by the AIA; we extend our sincere thanks for their generosity in sharing the survey with member organisations.
Author bio
Henry Wyard is a Senior Policy Analyst at Themis.
