Achieving effective compliance with automation throughout the customer journey

Financial institutions (FIs) have strong regulatory and financial incentives to address the $1.6 to $3.6 trillion in illicit funds circulating through the financial system annually. However, there is also a strong ethical dimension to this fight – behind the vast sums of money laundered are statistics like the 50 million people estimated to be living in modern slavery. 

Compliance teams and professionals, including Money Laundering Reporting Officers (MLROs) and compliance officers, are the “point of the spear” in preventing illicit activity, as our Global Head of FCC Strategy Andrew Davies put it in a session from our CATALYST event series. 

This article recaps Andrew’s discussion with Paul Simmons, MLRO at Hargreaves Lansdown, on how compliance teams can best be empowered with technologies like artificial intelligence (AI) to do their jobs effectively, levelling up FIs’ ability to tackle financial crime. 

The moral and commercial drivers for AI adoption in compliance

For regulated firms, mitigating risks from financial crime amounts to far more than compliance as a box-checking exercise. As Andrew and Paul discussed, it is a core commercial necessity tied to customer trust, and firms will face consequences beyond regulatory action if their anti-money laundering and countering the financing of terrorism (AML/CFT) compliance programs are inadequately designed. 

“The number one reason why consumers of financial services will leave a financial services company – and the flip side of that, why they will adopt a relationship – is because of trust, the bond of trust. People don’t want to be associated with organizations that are linked to financial crime, whether that’s fraud, or whether the organization is known to be providing financial services to criminals or terrorists.” 

Andrew Davies, Global Head of FCC Strategy at ComplyAdvantage. Hear more from Andrew by accessing the full CATALYST event on-demand. 

Compliance leaders therefore face a complex balancing act between maximizing efficiency to meet growth objectives, managing serious financial crime risks and protecting their customers, and adopting new technologies to combat increasingly sophisticated criminal actors who are, in many cases, are already exploiting AI themselves. Key drivers for technology adoption in compliance include: 

• Customer and victim protection: Safeguarding people from becoming victims of financial crime by swiftly detecting and acting on red flags is a constant priority for compliance teams. 
• Regulatory obligations: Firms must contend with increasingly demanding regulatory standards around customer due diligence (CDD), ongoing monitoring, and transaction monitoring, requiring the adoption of advanced technologies. Regulators themselves are increasingly backing AI-first solutions to detect and prevent financial crime. 
• Reputational risk management: With consumer trust and the reputation of the business on the line, manual processes and legacy systems often incur unnecessary risks for compliance teams. 
• Clearing the path to growth: Compliance tools that perform effectively while minimizing the amount of friction in the customer experience can help firms to scale by increasing their customer acquisition rates and transaction volumes. Meanwhile, according to some estimates, US companies alone lose around $3 trillion annually through having to deal with fragmented, missing, and outdated data, an issue which AI adoption and unified data architecture are essential to overcome.

Watch all Catalyst sessions on-demand

Catch up with all the keynote speeches, fireside chats, and panel discussions from Catalyst. Hear exclusive insights from our expert speaker lineup on-demand.

Watch now

Leveraging AI across the customer lifecycle

The financial crime risk management lifecycle encompasses every stage of the customer journey: from initial due diligence at onboarding, to ongoing monitoring of customer and transactional behavior, to case investigations and regulatory reporting. AI can offer crucial improvements to productivity and accuracy at every stage. 

1. Customer due diligence

Balancing secure onboarding with a seamless customer experience is a difficult balance for compliance teams to strike; customer abandonment rates for financial services products are sometimes estimated to be as high as 80%. 

AI can help speed up key processes for customer due diligence (CDD), including data ingestion and screening searches. More sophisticated solutions will also use AI to enrich data by extracting contextual insights, allowing compliance teams to go beyond simple name-matching and understand potential threats on a deeper level. 

2. Ongoing monitoring and transaction monitoring

Once a customer is onboarded, AI becomes instrumental in identifying anomalies based on your risk-based approach. AI can identify anomalies in customer behavior – for example, at an investment platform like Hargreaves Lansdown, a sudden shift from investing in long-term funds to more volatile assets, such as cryptocurrency, or behavior that suggests account takeover. 

Much as it does at the onboarding phase, AI can also constantly monitor risk data for changes – meaning any updates to customer risk profiles are immediately surfaced and reflected in dynamic customer risk scores. 

3. Alert remediation 

Undifferentiated alert backlogs pose significant challenges for compliance analysts, who may spend valuable time on lower-risk cases while leaving high-risk alerts unattended due to a lack of an effective way to triage alerts. Thanks to the development of advanced technologies like agentic AI, alert investigations can now also be automated at a level calibrated to your risk thresholds. 

AI agents are AI systems that can act autonomously without needing human intervention, independently pursue specific goals, and make context-aware decisions. In the context of AML investigations, AI agents can automatically assign themselves low-risk alerts, immediately remediate clear false positives (for example, due to country or name mismatches), and generate an auditable rationale for all decisions.

If there are elements of ambiguity to an alert and the agent is unable to discount it as a false positive, the case can be escalated for manual review. By streamlining the remediation process, AI agents free up human investigators to focus their expertise and judgment on cases that truly require their attention.

The conditions for successful AI adoption in compliance

Despite the massive productivity and accuracy gains that can be achieved through effective deployment of AI, over-reliance can become a significant pitfall. Understanding the capabilities and limitations of AI solutions, rather than treating them as a catch-all solution that can be simply layered over existing compliance programs, is essential to success. 

Compliance leaders must resist the pressure to adopt AI blindly or use their function to satisfy a generalized executive interest in AI solutions. Instead, effective AI adoption in compliance depends on: 

• Clearly defined use-cases: With applications from customer due diligence all the way through to case remediation, you should define your objectives and KPIs before initially adopting AI in specific, targeted ways. 
• Support for human expertise: While AI can excel at data analysis and pattern recognition, it lacks the compliance experience and expertise that allows human analysts to identify and deal with the most complex cases. AI tools should be regarded as a supplement to, not a replacement for, existing compliance teams. 
• Strong data and tech architecture: The success of any AI tool in AML depends on the data it operates on; even sophisticated solutions will be unable to fix problems caused by fragmented or missing data. Using AI as part of a unified platform, rather than as an overlay on legacy systems, is essential. 
• Robust oversight: Frequent testing and internal audits, a model risk management strategy, and prioritizing explainability are all critical in minimizing regulatory risks from AI use. 

“There are loads of opportunities with AI, which I think we’ll all embrace, but it’s not going to solve everything. Applied correctly, it will cut down on background noise and false alerts. It will help create efficiency so our teams can really focus on the value-add areas. So I see it as an area to enhance our controls.”

Paul Simmons, Group Head of Financial Crime & MLRO, Hargreaves Lansdown 

Unlock your compliance team with end-to-end AI solutions 

ComplyAdvantage Mesh is powered by AI at every stage of the financial crime risk management lifecycle to drive improvements in speed, accuracy, and explainability. Our AI capabilities include: 

• Initial due diligence: Data collection and curation is enhanced by AI for real-time, highly accurate results, with entity resolution to detect hidden links between individuals and companies. Advanced search techniques and probabilistic scoring further reduce false positives, while dynamic customer risk scoring aligns screening to your risk-based approach. 
• Ongoing due diligence and monitoring: Continuous data screening ensures that customer risk profiles are updated as soon as changes occur; alert prioritization enables the highest-risk cases to be addressed first. 
• Remediation and reporting: Our dedicated case remediation agent tackles low-risk alerts to clear backlogs and free up expert analysts for complex decision-making. 
• Underlying governance and efficacy: Mesh has been developed with principles of explainability, responsible AI, and model risk management in mind to ensure transparency and mitigate risks around AI bias and fairness.