Citi Payment Outlier solution launches in 90 countries
Citi’s Payment Outlier Detection solution has launched in 90 countries. Created by Citi’s Treasury and Trade Solutions business, it uses advanced analytics, artificial intelligence (AI) and machine learning (ML) to help proactively identify outlier payments – payments that do not conform to clients’ past patterns of payment activity – and allows clients to review and approve or reject such outlier payments via Citi’s institutional electronic banking platforms, CitiDirect BE and CitiConnect.
The bank says that client benefits of the service include:
- Enhanced control and monitoring of payments.
- Potential to reduce risk associated with outlier payments and subsequent losses.
- Advanced statistical machine learning algorithms instead of legacy rules-based logic to analyse payment patterns.
- Unique tailored customer profiles that identify individual payment patterns.
- AI-driven constant learning based on client usage.
- Real time alerts before outlier payments are sent to the beneficiary.
Cyber attacks have become increasingly frequent and sophisticated, and banks need to be more vigilant than ever, especially given increasing transaction volumes due to automation and digitisation. Client’s corporate treasury functions are also being challenged with these changes, and together with the rapid growth of instant payments, their expectations for speed in processing payments continue to rise.
The solution’s machine learning technology automatically adjusts controls to monitor discrepancies and changes in client payment behaviour, allowing for the analysis and identification of potential anomalies in affected payments before they are sent for clearing. It does this while helping to ensure that payments are processed quickly and efficiently.
This launch follows an extensive global pilot with 20 clients and is built on design principles that require minimum integration effort for clients. A key differentiator for Citi Payment Outlier Detection is its usage of advanced statistical machine learning algorithms instead of legacy rules-based logic to analyse payment patterns. This enables the system to automatically adjust in response to changing payment patterns as businesses evolve, expand and globalise. The solution also has client configurable product features such as the ability to synchronise payment release with the cut-off times, enhanced analytics and an increased choice of entitlement options.