Canadian firms can no longer rely on well-written policies to satisfy regulators. Since Bill C-12 received royal assent on 26 March 2026, amending the Proceeds of Crime (Money Laundering) and Terrorist Financing Act (PCMLTFA), the central question is no longer whether an anti-money laundering (AML) program exists but whether it works.
The legislation creates a new, very serious violation for failing to ensure a compliance program is “reasonably designed, risk-based and effective.” Penalties have risen forty-fold, with cumulative fines capped at the greater of C$20m or 3% of gross global revenue.
Six months on, Andrew Davies, Global Head of Financial Crime Compliance (FCC) strategy at ComplyAdvantage, and Claude Baksh, President and Co-Founder at Grace CSI, assessed the practical impact in a recent webinar.
The Financial Transactions and Reports Analysis Centre of Canada (FINTRAC) updated its administrative monetary penalty guidance in May 2026, changing how programs are judged. Claude said, “They’re no longer stopping at that evaluation of your written policies and procedures or your training. Now they’re looking at the operational effectiveness.”
The shift is already visible in enforcement. A notable share of penalties issued since March relates to repeated failures to file suspicious transaction reports (STRs). FINTRAC also benchmarks firms against peers with similar products and risk exposures, so a low STR count alone can trigger an examination. Claude Baksh said, “Your reporting output now serves as direct evidence for or against your program’s effectiveness.”
Efficiency and effectiveness are often confused, and alert backlogs expose the gap between them. A system generating more alerts than staff can clear is, by definition, ineffective. Andrew Davies raised a concern “If you’re overwhelmed with alerts, if you’ve got that huge operational backlog, how can you possibly be effective?”
Technology is a core weakness. ComplyAdvantage research found that 35% of Canadian firms report limitations in sanctions and watchlist screening, with respondents running an average of more than six separate screening tools. Fragmented legacy platforms, inconsistent data definitions, and off-the-shelf rules create noise rather than meaningful filings. Fixing this starts with unified data capture before thresholds are recalibrated.
Firms should be ready to provide evidence of three things: a documented risk assessment covering every live detection rule, logged justifications for threshold changes over time, and end-to-end STR trails showing why alerts did or did not result in filings. Where models inform decisions, explainability is essential. Claude Baksh added, “You’ve got to have notes on your files. You can’t just have automated decisions being made without that plain language explanation that’s factual, that you can defend.”
Securing budget may mean framing AML investment more broadly, supporting fraud detection and commercial insight. Andrew concluded, “Your AML system has historically got a lot of information about your customers and their behavior. Let’s look at that behavior through a different lens, and maybe there’s a business opportunity there.”
The firms best positioned are those that can trace the line from risk assessment to alert to filing, and explain every decision along the way.
Copyright © 2026 RegTech Analyst
Copyright © 2026 RegTech Analyst





