WhyAML study finds 49% of UK compliance questions need narrowing or referral

Go 2 Glory Ltd says an analysis of 1,587 real compliance questions from UK practitioners found nearly half could not be answered safely as asked. The company is launching WhyAML Companion to help firms classify questions, avoid tipping-off risks and keep a record supervisors can review.

Why it matters: - The study suggests general-purpose AI is unsafe for a large share of anti-money laundering queries because compliance answers often depend on missing facts, legal risk and firm-specific authority. - UK firms need a way to answer routine questions quickly without creating tipping-off risk or losing the audit trail supervisors expect. - Go 2 Glory Ltd is positioning WhyAML Companion as compliance-grade AI that can answer practitioner questions while preserving a defensible record.

What happened: - Go 2 Glory Ltd announced WhyAML Companion, a new feature inside the WhyAML practice portal for accountancy practices. - The launch is based on an analysis of 1,587 compliance questions from UK practitioners in 2026 across 135 subjects. - The study found 51% of the questions could be safely answered as asked, while 49% required narrowing, referral or refusal.

The details: - Of the 777 questions that could not be answered as asked, 303 needed more facts before a safe answer was possible. - Another 326 questions carried a tipping-off risk that changed what could properly be said. - 86 questions had to be referred to the firm's nominated officer. - 62 questions had to be declined outright. - WhyAML said the system classifies a question before answering it, so missing facts trigger follow-up, tipping-off risks limit the response, referral cases go to the nominated officer and prohibited questions are declined. - The platform uses a managed repository for periods, thresholds and deadlines, and shows each rule with the provision it came from plus who reviewed it and when. - WhyAML Companion keeps obligations and limits separate, so record-retention duties are not confused with retention ceilings. - The tool logs each exchange to the practice's own record, including the question, classification, authorities relied on, source version, outcome and timestamp. - The system is built into the WhyAML practice portal, so it can use firm background, obligations and nominated officer details already held there. - The practitioner's words stay inside the practice; the system stores only a redacted restatement with names removed and the key facts preserved.

Between the lines: - The findings show the core risk is not model quality alone. The same wording can be lawful in one context and criminal in another, especially where tipping off may be an offence under section 333A of the Proceeds of Crime Act 2002. - Go 2 Glory is arguing that compliance AI must understand firm context, not just answer fast. - The company also is targeting a common governance gap: firms often use consumer AI tools without formally documenting an adoption decision or risk assessment. - The product is designed to give firms documentation they can adopt as their own, which may matter as much as the answer itself. - Chae J. Pak, founder and CEO of Go 2 Glory Ltd, said practitioners are already using AI or forums for compliance questions but cannot show a supervisor what they asked or why.

What's next: - WhyAML Companion is available now inside the existing WhyAML practice portal at whyaml.com. - Go 2 Glory Ltd is offering the product as part of a broader compliance infrastructure platform for regulated businesses. - The company says the new tool extends WhyAML beyond client due diligence and into the practitioner's own compliance decision-making.

The bottom line: - The study's message is blunt: nearly half of real compliance questions cannot be answered safely without first classifying the risk, and the record of that decision may be as important as the answer.