It is true that insurance data is messy, customer touchpoints are limited, and behavioral signals are weak. When combined, this creates a perfect storm for high false positive rates in AML detection . Legacy systems, built on static rules, flag anything that might be suspicious but struggle to differentiate noise from real risk.
So, we tolerate it. We build large triage teams and add more layers of manual review. We slow down onboarding or claims processing, and we chalk it up to the cost of doing business.
But this is an incorrect way of looking at things. False positives are not inevitable and reducing them can significantly enhance your compliance efforts.
False positives don’t just create extra work – they erode the effectiveness of your entire compliance program. Here’s how:
According to a recent industry analysis by Datos Insights, "The AML models that many financial institutions use routinely generate 90-95% false positive rates ”, emphasizing the inefficiencies caused by outdated rules-based approaches.
That’s not just inefficiency. That’s systemic noise drowning out your compliance signal.
The insurance sector needs a smarter, more scalable solution, one that doesn’t rely solely on “if-this-then-that” rules.
That’s where AI-powered detection models come in. By learning from historical case outcomes, customer behavior patterns, and typology evolution, AI-enhanced tech can:
The result isn’t just fewer alerts, but better alerts. Financial crime teams gain the ability to focus their expertise where it truly matters - on the highest-risk, highest-impact cases, rather than being overwhelmed by noise.
Investigations become faster, more precise, and more defensible, supported by transparent models that regulators can trust and auditors can verify.
Ultimately, AI-driven detection empowers institutions to shift from reactive monitoring to proactive risk management. This strengthens compliance resilience while freeing resources to focus on innovation and customer trust.
With financial crime evolving and regulatory scrutiny rising, AML teams in insurance must make the shift from volume to value.
The future of AML in insurance won’t be defined by how much activity you review, but by how intelligently you detect what truly matters.
Accepting high false positives as “normal” leads to burnout, budget bloat, missed threats, and regulatory exposure.
Banks are ditching legacy rules in favor of adaptive models that cut through the false positives and improve detection accuracy. Insurers must follow suit or risk being left behind in the fight against financial crime.
Coming up next in our “Compliance myth-buster series: Insurance edition”:
Myth #3: “Rules are enough for AML” – Why static detection frameworks can’t keep up with dynamic criminal behavior.
Compliance myth-busters: Insurance edition: Myth #1: AML insurance—still low risk?
Compliance myth-busters: Insurance edition: Myth #4: If it's not regulated, it's not a risk
Compliance myth-busters: Insurance edition: Myth #5: AML and fraud teams can operate in silos
Redefining Risk: The Insurance Industry’s New Reality
Webinar: Regulators, risk & reinsurers: AML’s New Frontier
Data Sheet: Compliance for Insurance
Download our white paper “Elevating compliance in insurance: A risk-driven, AI-powered approach to AML and sanctions screening”.
False positives are common because legacy AML systems rely on static, rules-based detection - often designed for banking. These systems struggle to handle sparse customer data, limited behavioral signals, and fragmented insurance workflows, especially in non-life products. As a result, they flag large volumes of normal activity as suspicious without proper context.
Not necessarily. When 90–95% of alerts are false positives (as shown in industry benchmarks), investigators spend the majority of their time clearing non-issues. This means real threats may go undetected, or are investigated too late. High alert volume does not equal high detection quality.
Yes. AI can learn from historical case decisions, customer behavior, and known typologies to better distinguish true anomalies from normal variation. This enables fewer, more accurate alerts without compromising regulatory defensibility.
Not with the right solution. Modern AML platforms use explainable AI, which provides clear justifications for alerts, risk scores, and model behavior. This is critical for regulatory audits and internal trust. It's becoming a must-have under evolving global guidelines.
Accepting high false positives leads to:
Over time, it becomes a competitive disadvantage.
Thierry Fortin is a seasoned financial technology professional with over 25 years of experience in banking, consulting, and enterprise software implementation. Currently based in Paris, he serves as a key member of the SymphonyAI Financial Services sales team, where he supports the delivery and adoption of advanced AI-driven financial crime solutions. Prior to joining SymphonyAI, Thierry spent over a decade at BAE Systems Digital Intelligence, where he held roles including Senior Business Consultant and Solutions Consultant, working closely with leading financial institutions to implement risk and compliance technologies across Europe. His earlier roles include project positions at Société Générale and Crédit Foncier de France, where he specialized in risk management systems, securitization projects, and client risk detection solutions. He also brings international experience from his time in the U.S. with Nordstrom, where he managed logistics systems and led audit-related tech initiatives. With a deep technical background in systems development, business analysis, and project management, Thierry brings a unique combination of hands-on experience and strategic insight to every engagement. He is passionate about driving innovation in financial services, particularly in the fight against financial crime.
.svg)