Enhanced Due Diligence (EDD)

Standard due diligence tells you who the customer is. Enhanced due diligence (EDD) tells you whether you should be doing business with them at all. When a politically exposed person, a complex multi-jurisdictional corporate structure, or a customer in a high-risk country triggers EDD, the investigation can take weeks and cost thousands of pounds. Most of that time is spent gathering information, not making decisions.

What is enhanced due diligence?

Enhanced due diligence is the heightened level of investigation applied to customers who present an elevated risk of money laundering, terrorist financing, or other financial crime. It goes beyond standard customer due diligence in three dimensions: depth of investigation, breadth of information gathered, and seniority of approval required. Where standard CDD might verify identity and assess basic risk factors, EDD investigates the source of wealth, the source of funds for specific transactions, the customer's network of relationships, and the commercial rationale for the business relationship.

The triggers for EDD are defined by regulation and by the institution's risk appetite. The Money Laundering Regulations 2017 require EDD for politically exposed persons (PEPs), customers in high-risk third countries identified by the EU or the Financial Action Task Force (FATF), and correspondent banking relationships. Beyond these regulatory mandates, institutions apply EDD based on their own risk assessments: complex ownership structures, unusual transaction patterns, adverse media findings, or high-risk industry sectors.

The cost per case is high. Industry benchmarks put EDD at 5,000 to 15,000 pounds per case for complex corporate customers, with elapsed times of two to six weeks. The cost is driven by the manual research required: tracing beneficial ownership through multiple corporate layers, verifying source-of-wealth claims against public records, reviewing adverse media in multiple languages, and documenting the analysis to a standard that satisfies the regulator during examination.

The landscape

The FATF's grey list and the EU's high-risk third-country list determine which jurisdictions trigger mandatory EDD. These lists change periodically, and each change creates an operational wave: customers linked to a newly listed country must be identified, re-assessed, and potentially subjected to EDD. The practical challenge is identifying all affected customers across multiple systems and relationship types, which is itself a data integration problem.

The FCA has emphasised that EDD is not a box-ticking exercise. Enforcement actions have targeted institutions where EDD was performed but the analysis was superficial: boilerplate risk assessments copied between cases, source-of-wealth documentation accepted without verification, and approval decisions made by individuals who had not reviewed the underlying evidence. The regulator expects genuine, case-specific analysis proportionate to the risk.

PEP management remains operationally burdensome. The definition of a PEP extends to family members and close associates, which broadens the population significantly. Identifying whether a customer's spouse is a senior government official in a foreign jurisdiction, or whether a business partner holds a position that qualifies them as an associate, requires research that automated screening tools only partially address. False positives in PEP screening run high, consuming analyst time on cases that ultimately require no additional action.

How AI changes this

Automated source-of-wealth investigation is the application with the highest impact on EDD cycle times. AI systems that aggregate corporate registry data, property records, company financials, and publicly available wealth indicators can construct a preliminary source-of-wealth profile in minutes rather than days. The analyst reviews and validates this profile rather than building it from scratch. For cases where the source of wealth is straightforward (salary from a known employer, proceeds from a documented business sale), this can reduce investigation time by 60 to 70 per cent.

Corporate structure unravelling uses entity resolution and graph analytics to trace beneficial ownership through layered corporate structures. An AI system that can follow the chain from a customer entity through holding companies, trusts, and nominee arrangements to the ultimate beneficial owners, cross-referencing against corporate registries across jurisdictions, accelerates the most time-consuming element of corporate EDD.

Adverse media screening enhanced by natural language processing covers more sources in more languages with greater precision than keyword-based approaches. The system distinguishes between a customer who shares a name with a person mentioned in negative news and a customer who is the subject of that coverage. Contextual analysis reduces false positives and ensures genuine adverse findings are not buried in noise.

Risk-based EDD prioritisation uses ML models to score which EDD cases are most likely to result in a decision to exit the relationship or to file a suspicious activity report. Prioritising these cases ensures that the highest-risk customers receive attention first, rather than cases being processed in the order they entered the queue. This is particularly valuable when EDD backlogs accumulate, as they do after regulatory changes or list updates.

What to know before you start

EDD automation does not mean EDD without humans. The regulatory expectation is that a qualified professional reviews the evidence and makes the risk decision. AI accelerates the evidence gathering and preliminary analysis. The decision remains human. Design your system with this boundary clearly defined, and ensure your audit trail captures both the AI-generated analysis and the human judgement.

Data availability varies by jurisdiction. Automated source-of-wealth investigation works well in countries with accessible corporate registries and property records. It works poorly in jurisdictions where these records are unavailable, unreliable, or behind paywalls. Assess the data landscape for your highest-volume EDD jurisdictions before committing to an automated approach.

PEP screening accuracy depends on the quality of the underlying PEP database. Commercial PEP databases vary in coverage, currency, and accuracy. A system that screens against an incomplete database provides false assurance. Evaluate your PEP data provider's methodology, update frequency, and coverage of the jurisdictions relevant to your customer base.

Start with the EDD backlog. Most institutions have a queue of overdue periodic EDD reviews. Automating the evidence-gathering step for these reviews delivers immediate value, clears the backlog, and provides a controlled environment to validate the AI's output against established analyst expectations. Once validated, extend to new-to-bank EDD and transaction-monitoring triggered EDD reviews.