The centre of gravity is shifting from static onboarding data to live behavioural evidence, redefining how risk is detected and managed. In a recent discussion between Carsten Helm and Oonagh van den Berg, in collaboration with AML Watcher, Transform Finance, and RAW Compliance, the focus turned to how financial crime is evolving at a pace that outstrips most existing control frameworks.

Below, we explore how this shift is driving transformative and systemic change across global GRC frameworks in the financial industry.

• From Static Profiles to Behavioural Truth

For decades, KYC (Know Your Customer) has been seen as the bedrock of compliance. But in a world of complex financial crime, its limitations are stark. Static profiles capture only what a customer tells you at onboarding — often outdated, incomplete, or deliberately misleading.

Carsten Helm highlights the gap:

“KYC measures are limited in effectiveness. You have to believe what the customer is saying. Unless you bring in more information from transactions and machine learning, you don’t get the full picture.”

Oonagh van den Berg takes it further:

“Without digital identity, KYC isn’t the value-add we think it is… it’s not until they start operating the account that we actually find out who they are.”

Why It Matters

• Data waste is staggering. Onboarding typically generates ~1,400 data points per customer, yet banks operationalise only 120–140 on a day-to-day basis. Over 1,000 datapoints sit unused — leaving enormous intelligence value untapped.

• Static checks miss real risk. Traditional KYC provides a snapshot in time. Behavioural telemetry — transaction patterns, device usage, counterparty networks — offers a higher-fidelity, continuously updated risk signal.

• Industry leaders are already shifting. HSBC has publicly disclosed screening ~1.2 billion transactions per month, leveraging AI to reduce noise and increase detection precision (Google Cloud, 2023).

• Academic and trade evidence agrees. Studies consistently show legacy rules engines producing false-positive rates over 90%, while modern ML techniques (entity resolution, network graph analytics) significantly improve both detection and efficiency (ScienceDirect; Retail Banker International).

What Good Looks Like

• Treat Transaction Monitoring as “continuous KYC.” Behavioural data should flow back into profiles, constantly refreshing risk ratings.

• Leverage dynamic features. Incorporate velocity indicators, counterparty/merchant clustering, device and geo heuristics, and graph centrality scores into customer risk models.

• Fuse external signals. Enrich with adverse media, sanctions lists, and shared industry alerts to spot anomalies faster.

• Close the feedback loop. Feed high-signal insights into onboarding, EDD, and periodic reviews — ensuring profiles reflect reality, not just static declarations.

The Bottom Line

Static KYC will always have a role, but it is no longer enough. Real risk emerges from behaviour, not declarations. Institutions that treat transaction monitoring as “continuous KYC” and operationalise their full data estate are the ones truly moving towards behavioural truth in compliance.

WATCH THE FULL PODCAST BELOW