Zenus Bank announced on June 11, 2026 that it had received Celent's Model Risk Manager Award for Operational Resilience. The award issuer's assessment centers on a specific operating design: connecting checks on customers, institutions and transactions so risk decisions can be made while banking services are being delivered. That is a consequential design choice for a bank supplying infrastructure behind other companies' brands.
For an institutional banking partner, those controls affect the product directly: the time needed to admit a customer, the handling of a payment exception and the information available when activity requires investigation. Zenus's approach gives the bank a shared way to coordinate those tasks across payment routes as its embedded business expands.
Chief Compliance Officer Gabriel Viera described the approach as embedding compliance within the banking experience. His explanation places the bank's risk controls alongside the accounts and payments it supplies, so institutional growth and regulatory oversight are managed within the same operating framework. Source announcement.
An orchestrator makes separate controls work together
The architectural detail is set out in Ian Watson's June 11 Celent case study. It identifies Tuum as the core banking engine, Exiger for institutional due diligence and NICE Actimize for monitoring and case management. Identity verification, blockchain analysis, settlement and card connectivity also form part of the environment. Zenus's own orchestration and risk layer connects those components.
That coordination addresses a concrete problem. A partner's risk profile can change after onboarding; a transfer may need to be assessed alongside activity on another rail; a customer record may need to be available to an investigator who is reviewing an alert. If the systems cannot share the relevant context, staff have to reconstruct it at the point when a decision is needed.
A common decision environment can preserve the relationship between the customer, the account and the activity being reviewed. Its value is therefore broader than automation of a single check. It can make it easier to see why a payment was released, escalated or stopped, and to connect the investigation with the information available when the transaction occurred.
The implementation was also a coordinated staffing effort. Celent describes a 38-person core team comprising 28 compliance and financial-crime specialists and ten technology specialists, with Compliance, Engineering and Product working under shared governance. Celent case study, page 8.
The implementation brought together 28 compliance and 10 technology specialists
Celent describes a 38-person core project team working across Compliance, Engineering and Product.
people in the project team
Tap or hover for details. All values are also available in the data table.
Source
The team composition is reported in the June 11, 2026 case study. It is not a current company-wide headcount or a measurement of employee productivity.
Show data tableHide data table
| Category | Core implementation team |
|---|---|
| Compliance and Financial Crime | 28 people in the project team |
| Technology | 10 people in the project team |
A processing platform with a substantial review operation
The full study describes more than $4 billion in monthly payment volume and a $75 billion annualized run rate. It also reports 6,000-plus monthly investigations managed by a 28-person compliance team, partner onboarding reduced from up to six months to under a week, and $2 million to $3 million in client fraud losses recovered. The operating tables identify Zenus Bank as their source.
Those measures describe different dimensions of the operation. Investigations are work performed; onboarding time is a service cycle; recovered money is an outcome after loss. Without common reporting periods and denominators, they should not be collapsed into a single efficiency score or used to calculate a platform-wide fraud-loss rate.
Show data tableHide data table
| Metric | What is being measured | Operational role |
|---|---|---|
| Payment-volume run rate | Payment value expressed at an annualized pace | Describes processing scale; separate from revenue |
| Monthly investigations | The volume of compliance review work | Describes review activity rather than confirmed offences |
| Recovered fraud losses | Money recovered for affected clients | Describes a recovery outcome rather than a percentage loss rate |
| Partner onboarding time | The elapsed admission process reported by the bank | Connects institutional due diligence with access to the platform |
The network figure also needs a careful label. Across the study, relationships are described as institutions and as financial institutions, fintechs and corporate partners. Calling them all banks, or treating them as a count of retail customers, would introduce precision the source does not support.
From alerts to suspicious activity reports
Celent reports that suspicious activity report filings increased from 100 annually to more than 200 per month. It gives 2025 conversion rates of 16.2% from alerts to cases, 46.3% from cases to SARs and 7.5% from alerts to SARs. These use different denominators: all generated alerts, opened cases and filed reports are successive stages, not interchangeable populations.
Two of the published conversion measures can be placed on a common scale: 16.2% of alerts became cases and 7.5% of alerts led to a SAR. The case-to-SAR percentage describes a different stage of the process and retains its separate case-based denominator. Celent conversion metrics, page 10.
16.2% of alerts became cases; 7.5% led to SARs
Both measures use alerts as the denominator, allowing the two documented stages to be compared on the same scale.
%
Tap or hover for details. All values are also available in the data table.
Source
Bank-reported conversion rates for 2025, reproduced by Celent. The separate 46.3% case-to-SAR rate uses cases as its denominator and is not plotted. SARs identify suspicious activity; these percentages do not measure proven crime or crime reduction.
Show data tableHide data table
| Category | Share of alerts |
|---|---|
| Alert-to-case conversion | 16.2% |
| Alert-to-SAR conversion | 7.5% |
The progression describes the bank's investigative funnel: alerts are reviewed, a portion become cases and some cases lead to a report. Conversion rates can be affected by triage, risk mix and reporting practices. They help describe the operation without establishing a percentage reduction in financial crime.
FinCEN's guidance on suspicious activity reporting explains the division of responsibility: institutions identify and report suspicious activity, while investigating the underlying suspected crime remains a law-enforcement task. SAR output is useful evidence of a reporting operation. It is not a count of convictions, proven offences or customers found guilty of wrongdoing.
The July AI expansion addresses the investigator's starting point
A later NICE Actimize account, published July 1, 2026, describes Zenus adding InvestigateAI and NarrateAI to ActOne. The operational aim is to bring together customer, transaction, fraud and compliance information so investigators can begin with the relevant context, instead of gathering it manually across multiple systems.
The vendor lists shorter investigations, more complete risk assessment and greater capacity without proportionate staffing increases as intended benefits. It publishes no measured percentage improvement for Zenus. That distinction matters: a named implementation is stronger evidence than a general claim about AI, but it is still different from a before-and-after outcome study.
For a bank, a useful investigation narrative also needs to remain traceable to the records that support it. An apparently convincing summary is insufficient if a reviewer cannot establish which transactions, customer information and risk signals informed the conclusion. The commercial value of assistance depends on preserving the basis for the decision.
Operational resilience includes people and responsibilities
The regulatory context is substantial even when services are delivered through partner platforms. FinCEN's 2020 final-rule announcement for banks lacking a federal functional regulator extended minimum AML-program, customer-identification and beneficial-ownership requirements to those institutions. The absence of a federal functional bank supervisor does not mean the absence of federal financial-crime obligations.
More broadly, the banking agencies' 2024 industry statement on third-party deposit arrangements discusses access to records, clear contractual responsibilities and contingency planning. These concerns explain why coordination matters beyond the software itself: responsibilities and the information needed to discharge them have to remain usable when a problem occurs.
For incentive, prepaid and cross-border payout programs, that lesson reaches the end-user experience. A control environment has to distinguish a legitimate recipient awaiting payment from activity that requires escalation, and support a defensible response when a transaction is disputed. The Celent award recognizes a design intended to make such coordination possible at scale. Sustained performance will depend on the quality of decisions and service as well as the volume handled.
In this article
Daniel Brooks
Daniel Brooks covers payment networks, financial technology, payment security, gift card fraud and PCI DSS for Gift Card News.