Egmont Group ¦ The Role of FIUs in the Preparation of Strategic Intelligence

Strategic intelligence is where FIUs turn suspicious transaction reports into a national risk picture

Strategic analysis is one of the two analytical functions that FATF Recommendation 29 assigns to financial intelligence units (FIUs), and often the one that gets squeezed. A survey by the Egmont Group’s Information Exchange Working Group (IEWG), based on responses from 57 FIUs in Africa, Europe, the Americas, the Middle East, Asia-Pacific and Eurasia, maps how FIUs produce strategic intelligence: under which mandate, with how many staff, with which data and tools, and with what effect on supervision, policy and the national risk assessment (NRA). Every responding FIU is permitted to do strategic work, but capability varies widely: most remain at a “developing” stage, and only a small group runs a fully institutionalised, data-driven strategic function.

This matters for reporting entities as well. The typologies, red flags and sectoral risk assessments that banks, fund managers and other obliged entities build into their monitoring are strategic intelligence products, and their quality depends on the capacity the survey describes.

Strategic and operational analysis are different disciplines

Operational analysis starts from a suspicious transaction report, a foreign request or a law enforcement inquiry and asks who, what, when, where and how. It is case-driven, short-cycle and largely deductive, and it succeeds when it moves an investigation, a prosecution or an asset recovery forward. Strategic analysis looks at a phenomenon, a sector, a method or a payment corridor. It aggregates hundreds or thousands of cases, reasons inductively and abductively with statistical, typological, geospatial, network and machine-learning methods, and produces products meant to stay relevant for a year or more. Its readers are policy ministries, supervisors, the NRA process, international bodies, the private sector and FIU leadership, and its success shows in policy decisions, supervisory priorities and the quality of the national risk picture.

Mixing the two causes predictable problems. Strategic work treated as a residual of operational work never gets done, because operational demand is always urgent. Operational analysts asked to deliver strategic products without protected time tend to produce extended case summaries instead of pattern intelligence. Strategic products written for operational readers fail to reach policymakers, who need different framing and evidence. A mature FIU therefore protects strategic capacity from operational displacement while keeping both functions connected: casework supplies the empirical base, and strategic findings sharpen typologies and the indicators used to triage reports.

Bastian Schwind-Wagner _Follow

"Strategic analysis is where an FIU turns thousands of reports into knowledge that supervisors, policymakers and reporting entities can use. The Egmont survey shows that almost every FIU is allowed to do it, but most still do it with small teams, hybrid roles and limited access to the data that give suspicious transaction reports their meaning.

For compliance teams, the link is direct. The typologies and red flags that feed our monitoring systems are only as good as the strategic capacity behind them, and defensive or low-quality reports make that work harder. Better reports and closer use of FIU feedback are the private sector’s contribution to a sharper national risk picture."

Every FIU may do it, but only half have it defined in law

All responding FIUs say their legal or institutional framework permits strategic analysis. Only about 53% report that their AML/CFT legislation actually defines it, and roughly one-third operate without explicit statutory authority, relying on broad analytical powers or institutional practice. Without a clear anchor, strategic analysis competes with casework, depends on the commitment of individual leaders and is exposed when priorities shift. A clear legal basis also supports resourcing, data access and the credibility of strategic products with policy stakeholders, and it protects the FIU when it reports on sensitive sectors or institutional weaknesses. Closing this gap costs little in budgetary terms.

Where definitions exist, they mostly follow the Interpretive Note to Recommendation 29 closely. That helps peer review and cooperation, but it also makes them largely descriptive and retrospective, focused on trends, patterns and typologies seen in past cases. Forward-looking functions such as anticipatory risk modelling, horizon scanning and scenario analysis exist in practice in some FIUs, often through technology projects, academic partnerships or public-private partnerships, yet they are rarely written into law. Since budget discussions and legal assessments of data use tend to follow the wording of the mandate, the omission has practical consequences.

Small teams, hybrid roles and a skills gap in data science

Around 81% of responding FIUs have a dedicated strategic analysis team, but these teams are small: a median of five full-time equivalents and about 10% of the workforce, against roughly 40% for operational analysis. Nearly half of strategic analysts (47%) also do operational work. Hybrid roles keep strategic work grounded in operations, but urgent casework tends to push longer-cycle projects aside, especially in smaller FIUs. The proposed answer is either a protected core of dedicated analysts or, for smaller FIUs, ring-fenced strategic time backed by leadership.

Self-assessed skills follow a clear pattern. Critical thinking (4.0 out of 5) and policy analysis (3.7) score highest, risk forecasting (3.5), data science (3.3) and econometrics (3.3) lowest. Training shows the same imbalance: most FIUs have received training in strategic methodology (79%), report writing (59%) and critical thinking (55%), but only 40% in big data and about a quarter in machine learning. Where data science or analytics teams exist, their median size is three staff.

Data access: STRs are not enough

Suspicious and other transaction reports are available to all responding FIUs, and open-source intelligence to almost all. Access thins out for the data that put those reports in context. Beneficial ownership registers are available to 77%, supervisory data to 68% and academic research to 58%. Among administrative sources, company registers are near-universal (93%), while passenger and travel data reach only 63%. Customs and trade data (86%) and law enforcement information (81%) are more widely available but not consistently integrated.

Strategic analysts are less troubled by the number of STRs than by missing context. Defensive, duplicated or low-quality reports also add noise and make strategic patterns harder to detect, which concerns reporting entities directly. Closing the data gaps needs structured agreements with company registries, tax and customs authorities and foreign FIUs; FIUs cannot do it alone.

Tools, typology libraries and outdated detection rules

Technical capability scores just below the midpoint (2.9 out of 5), innovative practices such as big data, AI and OSINT lower still (2.5), and 32 of 57 FIUs say their systems are not adequate for advanced strategic analysis. Toolsets range from goAML, Excel and BI dashboards to SQL, Python and network analysis tools, with a few FIUs running machine-learning models, data warehouses and cloud platforms. The tools most in demand are natural language processing for STR narratives, large language models, graph analytics, scalable data integration and automated anomaly detection.

About 77% of FIUs see a need to recalibrate their machine-learning models or detection rules, a sign of weak model lifecycle management. Some 68% keep a typology library, but many update it ad hoc or after events instead of on systematic review cycles, which limits its use for prediction and prioritisation.

What FIUs produce and who uses it

Typology reports are close to universal, followed by trend studies, strategic assessments, risk indicators and red flags, and advisories. More advanced FIUs add national and sectoral risk assessments, dashboards and outputs built with data science. Around 90% disseminate to reporting entities, about 70% each to law enforcement and supervisors, about 60% to government, and only around 30% to the public and 20% to international partners. Most FIUs (86%) report some feedback loop and 81% run awareness sessions, yet few measure whether dissemination improves reporting quality, policy or operational outcomes.

Impact scores are highest for the contribution to NRAs (4.3 out of 5), and 58% of FIUs say their strategic analysis is extensively integrated into the NRA. Policy influence (3.7) and forward-looking risk identification (3.6) follow, while effectiveness measurement (3.2) and engagement in regional or bilateral threat assessments (3.0) score lowest. Seventy percent of FIUs report documented policy or regulatory changes resulting from their strategic work, including legislative amendments, supervisory guidance and revised national strategies. The strongest NRA integration comes with formal inter-agency protocols, co-authorship with the NRA coordinator and dedicated resources.

Strategic analysis is covered by FIU independence

Because strategic products are visible, policy-relevant and often built from anonymised or aggregated data, they are sometimes assumed to fall outside the independence that protects case-level analysis. The IEWG rejects that view. The Interpretive Note to Recommendation 29 protects “analysis” as a whole and does not distinguish operational from strategic work. Independence concerns analytical autonomy – who decides what is studied, how and with what conclusion – and has nothing to do with the sensitivity of the data. A strategic report written from fully anonymised data under outside instruction is, on this reading, more problematic than case analysis on identifiable data conducted under the FIU’s own direction.

Three pressures are named. In topic capture, policymakers dictate the strategic agenda; in conclusion shaping, drafts are revised for political acceptability; in publication control, decisions on whether and when to publish neutralise inconvenient findings. None of them requires bad faith. Coordination remains legitimate: the FIU contributes its findings to the NRA, national strategies and multi-agency typology work, which may pose specific questions and use negotiated language, provided the FIU’s own findings stay its own and are represented accurately. The FIU provides input, without the power to bind policymakers. Statutory clarity, documented methods and authorship, a pre-agreed publication policy such as a committed annual typology report, and benchmarking against the Egmont Group, the FATF and regional bodies help hold that balance.

Four maturity levels and what moves an FIU up

FIUs fall into four indicative maturity levels. Roughly 15–20% are “foundational”, with little dedicated capacity, weak data access and few formal outputs. Around 40–45% are “developing”, with hybrid roles, basic typologies and limited tools. About 30–35% are “established”, with dedicated teams, regular outputs and strong NRA integration, and only 5–10% are “advanced”, with AI and machine-learning tools, integrated data and measurable policy impact.

The correlations point the same way. An explicit legal mandate goes with higher and more regular output, dedicated teams with more diverse and sophisticated products, broader data access with deeper and more forward-looking analysis, and structured feedback with documented policy impact. The figures should be read with caution, since they are composite analytical estimates and not every Egmont member responded.

Recommendations and a second phase

The recommendations follow the findings. Jurisdictions should give strategic analysis an explicit mandate and a statutory definition, using secondary instruments such as FIU strategic plans and internal regulations where legislative reform is slow, and should reflect mutual evaluation findings in reform agendas. FIUs should set up dedicated units with protected staffing and separate reporting lines, while the Egmont Group should develop a competency framework and advanced training in data science, network analysis and machine learning, with priority for lower-capacity FIUs. Formal gateways to beneficial ownership, tax, customs, vehicle and land registers, shared technical experience and a possible shared analytics platform for smaller FIUs address the data and tool gaps. FIUs should also publish annual strategic intelligence reports, track the outcomes of their products through structured feedback, consider publishing suitably classified products and adopt a KPI framework. Governments should formalise the role of FIU strategic intelligence in the NRA, supported by a model NRA integration protocol.

A second phase would move from diagnosis to practice, with a typology of strategic products and quality standards, methodological guidance tiered by maturity, a compendium of good practices – including how FIUs shield strategic units from operational displacement and balance independence with multi-agency coordination – and practical templates, KPI sets and training modules.

What reporting entities can take from this

The typologies and red flags obliged entities receive are only as good as the strategic capacity of their FIU, which in most jurisdictions is still developing. Their own reporting counts at the strategic level too, since defensive or low-quality STRs blur the patterns analysts are trying to detect. FIUs with structured feedback and published strategic reports give compliance teams a better basis for calibrating monitoring scenarios and business-wide risk assessments. Groups operating across several jurisdictions can treat FIU maturity as one indicator of how far to rely on local typologies and where to add their own analysis.

Dive deeper

• Egmont Group ¦ The Role of FIUs in the Preparation of Strategic Intelligence ¦ Link