The UK's failure-to-prevent fraud offence is often described in the language of corporate criminal law: a new basis for liability, a new exposure for large organisations, a new incentive to tighten controls. That description is accurate but incomplete. The offence's practical centre of gravity lies in the organisation's own records. Where the statutory defence turns on whether reasonable fraud prevention procedures were in place, what ultimately matters is whether the organisation can prove it. A programme that exists in policy documents but cannot be reconstructed from contemporaneous evidence offers limited protection.
For legal and compliance teams, that reframing changes priorities. The task is not simply to adopt a fraud prevention framework, but to build a record that fits together: risk assessments that identify the exposures, due diligence files that show how counterparties were checked, training logs that demonstrate reach and frequency, monitoring reports that show controls were tested, and board minutes that show decisions were taken and followed up. Each element is familiar; what is less familiar is the expectation that they form a coherent evidentiary chain.
The offence and the defence
The offence was introduced by the Economic Crime and Corporate Transparency Act 2023 and came into force on 1 September 2025. It applies to large organisations, defined by reference to the Companies Act 2006 size thresholds. Where an associated person commits fraud with the intention of obtaining a benefit for the organisation, the organisation may be liable, regardless of whether senior management knew of or approved the conduct.
Liability does not depend on proving that the organisation was negligent, but the offence is not absolute. A statutory defence is available where the organisation can show that it had reasonable fraud prevention procedures in place, or that it was not reasonable in all the circumstances to expect it to have any. That defence is where the practical action lies. It places the organisation's own arrangements at the centre of the case and invites scrutiny of whether they were proportionate to the risks the organisation actually faced.
Two features of the offence deserve particular attention. First, the concept of the associated person is broad, reaching employees, agents, subsidiaries and other persons performing services for or on behalf of the organisation. Second, the benefit element means that conduct need not have enriched the organisation in fact; the intention to benefit is what counts. Fraud risk arising through intermediaries, suppliers, distributors and other third parties therefore falls squarely within the frame.
From policy framework to evidential record
Most large organisations now have an anti-fraud policy, a code of conduct, a whistleblowing line and a suite of compliance training modules. These are necessary, but they are not the same as an evidential record. The distinction is between describing a control and demonstrating that it operated.
A defensible record typically has several interlocking parts:
- Risk assessment. A documented assessment identifying where fraud could occur, how, by whom, and what controls respond to it — dated, owned, refreshed, with changes explained.
- Due diligence files. What was checked before a relationship began and during its life: ownership and control information, adverse media, sanctions and litigation checks, financial information, and the reasoning behind risk ratings.
- Training logs. Who was trained, on what, when, and with what completion rates. Training directed at high-risk functions carries more weight than generic annual modules.
- Monitoring and testing reports. Control testing, exception reviews, analytics on payments and expenses, and the conclusions drawn from them.
- Board and committee minutes. Evidence that fraud risk was considered at a senior level, that decisions were taken, and that actions were assigned and tracked.
- Escalation records. What was raised, by whom, when, how it was investigated, and what was done as a result.
The evidential value comes from the connections between these items. A risk assessment that identifies third-party payments as a high risk, followed by due diligence addressing that risk, training for the relevant team, monitoring that tests the control, and a board minute recording the outcome, tells a story. Documents that do not reference one another tell a much weaker one.
Third parties and contractual controls
Because associated persons can include intermediaries and other third parties, contracts become part of the fraud prevention record. Anti-fraud and anti-bribery representations, obligations to maintain controls, audit and information rights, notification duties where concerns arise, and termination rights for breach serve a dual purpose: they shape behaviour, and they create the means to detect and respond to problems.
Contract terms alone prove little. The record that matters includes documented supplier due diligence, the rationale for the level of scrutiny applied, evidence that contractual controls were exercised, and records of issues identified and resolved. An organisation that holds audit rights but has never used them, and cannot explain why, may find the clause of limited assistance.
Governance, testing and readiness
The defence is strengthened by visible ownership: a named board member or committee responsible for fraud risk, a clear allocation of responsibilities across legal, compliance, finance, procurement and internal audit, and a reporting line that reaches the board.
Testing is where many programmes are thinnest. Policies are refreshed annually; controls are rarely tested with the same discipline. A 2026 programme should include scenario-based testing of how the organisation would respond to a suspected fraud, remediation tracking with owners and deadlines, and investigation records that set out scope, methodology, findings and actions. These records serve a compliance purpose and a litigation purpose at once.
Readiness also has a documentary dimension. Retention schedules should ensure that risk assessments, due diligence, training records and investigation files survive long enough to be produced. Privilege should be considered before an investigation begins, not after, since the position will depend on who commissioned the work, for what purpose and how it was communicated. The organisation should also be able to explain its fraud prevention decisions: why a risk was rated as it was, why a control was designed as it was, and why a change was made.
Conclusion
The failure-to-prevent fraud offence is best understood as a test of institutional memory. The statutory question is whether reasonable procedures were in place; the practical question is whether the organisation can show it. That requires more than a refreshed policy suite: it requires risk assessments, due diligence, training, monitoring, governance and escalation records that reference one another and can be produced on demand.
For 2026, the more useful exercise is not another policy review but a test: pick a significant fraud risk, trace it from assessment through control design, training, monitoring and board oversight, and ask whether the resulting file would persuade an independent reader. Where the answer is no, the gap is evidentiary, and it can be closed before it is tested.
