Executive summary
AI adoption in anti-financial crime (AFC) is expected to rise rapidly despite limited current deployment. While 69% of organisations do not yet use AI, 51% plan to adopt it within three years. Technology remains the largest obstacle (39%), along with concerns around reliability, privacy, costs, internal knowledge gaps and regulatory uncertainty. Even so, most organisations anticipate accelerated adoption as tools and regulation mature, and 85% of leaders report at least intermediate understanding of AI’s risks and opportunities.
Background
Financial-crime recovery rates remain extremely low: Europol estimates that less than 2% of illicit proceeds are recovered annually, despite approximately USD 85 billion spent on compliance each year in EMEA. Existing AFC systems are often inefficient and fragmented. AI is increasingly viewed as a means to improve due diligence, risk detection, investigations and asset recovery. This study examines senior leaders’ perceptions of AI’s role in AFC.
Methodology
The survey, conducted from October 2024 to February 2025, gathered 74 responses from senior leaders across public, private and non-profit sectors, including CEOs, directors, compliance heads, MLROs and operational staff. The nine-question survey collected quantitative and qualitative insights on technology challenges, AI use, perceived value, risks and desired capabilities.
Key finding 1 – technology challenges persist, but AI adoption is imminent
Technology was the most frequently cited challenge (39%), followed by issues relating to efficiency, costs and change management. Legacy systems were described as inaccurate or inefficient. Current AI use remains limited, but adopters apply it across KYC, due diligence, transaction monitoring, adverse media, customer risk scoring, investigations and fraud detection. Planned adoption is strong: 51% of non-users expect to implement AI within three years and over 80% within five.
Key finding 2 – senior leaders recognise AI’s potential cross AFC practices
Around half of respondents reported that leadership understands AI “somewhat well” or “extremely well.” Efficiency is the primary driver, with over 70% seeking time and cost savings; around 45% aim to address emerging risks. Desired capabilities include automated onboarding workflows, behavioural risk assessment, real-time monitoring, cross-system integration and document-handling automation, with strong emphasis on explainability and interoperability.
Key finding 3 – despite barriers, clear pathways to adoption exist
Primary concerns include reliability (22%), high costs (18%), privacy and security (10%) and internal knowledge gaps (45%). Over one-third cited regulatory uncertainty as a challenge, though many expect this to improve. Training, hybrid human–AI models and improved product suitability are viewed as critical enablers.
Conclusion
AI adoption in AFC is expected to accelerate significantly over the next one to three years. Organisations should strengthen internal capabilities and proactively address regulatory, cost and knowledge barriers to prepare for AI-enabled AFC.
Recommendations
Recommended actions include adopting integrated AFC platforms, implementing real-time monitoring, auditing processes, integrating AI via APIs, expanding data sources and investing in training.
How themis uses AI
Themis Search applies AI to screening, monitoring, investigations and data harvesting, using NLP for media analysis, multilingual search and an AI chatbot for user support.
Themis UK research and innovation grant project
In 2024, Themis received a UKRI grant to develop AI that identifies hidden financial-crime relationships using real crime data and integrated investigative workflows.
About themis
Themis provides AI-enabled AFC solutions combining analytics, automation and proprietary intelligence to help organisations detect and manage financial-crime risks.
