UiPath Documentation
industry-department-solutions
latest
false
  • Overview
    • About financial services solutions
  • Financial Crime Compliance (FCC) solutions
UiPath logo, featuring letters U and I in white

Financial Services Solutions user guide

Last updated Apr 3, 2026

Evelyn - Name Screening Alert Review

As part of the onboarding and KYC process, financial institutions screen individuals and entities, including beneficial owners and related parties, as well as securities, for potential exposure to sanctioned entities, individuals, and securities posing a potential risk. Screening primarily occurs on the government-issued sanctions lists, such as the Office of Foreign Assets Control (OFAC), but it can also include other domestic and international watchlists including lists related to politically exposed persons (PEPs).

Financial institutions use sanctions screening or watchlist filtering technologies to compare input names and other data with items on a sanctions list. The outputs of those technologies are alerts and hits. Evelyn, the Name Screening Alert Review agent, is designed to ingest alerts and associated hits generated by a name screening system and determine whether they should be treated as false positives by examining the available input and watchlist data.

The agent supports multiple input sources and integrates with supported screening systems to work alongside the existing screening platform. Some financial institutions also use the agent to identify alerts that need more information and send them for manual review to a predefined team within the institution. In addition to determining if an alert is a false positive, Evelyn, the Name Screening Alert Review agent, creates a short decision narrative for audit trail and justification purposes.

Tip:

Watch a quick demo of how Evelyn, the Name Screening Alert Review agent, works!

High-level processing flow

At a high level, Evelyn, the Name Screening Alert Review agent, processes alerts through the following steps:

  1. Ingestion: retrieves data from a supported screening system, API input, or an uploaded CSV file and ingests alerts, hits, input information, and watchlist entity data. Connections to screening systems are typically established through an API.
  2. Data parsing: organizes ingested input names and associated demographic information into defined fields, including first name, last name, address, country, citizenship, date of birth, and ID number. This step parses watchlist entity data into name, entity details, and other relevant identifiers provided by an external or internal bank source.
  3. Data enrichment (optional): connects to sources outside the user's name screening system to obtain additional data to assist with alert review.
  4. Model: executes the classification model and related decision logic to assist in the false positive decision-making process at the hit and alert level. This process step is the primary driver of the false positive decision-making process.
  5. Decision and justification: assigns "false positive" or "needs more information" to each hit and alert. This includes creating a human-readable justification for each decision at the hit or alert level.
  6. Screening system/ case manager: connects to a screening system or case manager to submit hit and alert decisions together with the decision narrative. This connection is typically established through an API.

Key agent components

Evelyn, the Name Screening Alert Review agent, is composed of several core components that support your workflows:

  • Connectors: connects to sanctions screening systems through supported integrations, primarily APIs. Productized connectors to supported systems are available out-of-the-box.
  • Alert Parser: based on the format of incoming alerts and hits from the screening system, identifies and parses key inputs for the models and watchlist entity.
  • Configuration: UI-based configurations for rules, input type, model selection, data enrichment options, and output.
  • HITL interface: graphical UI that presents the adjudicated messages and alerts to the user and enables them to review or modify the decisions made by the model.
  • Analytics: enables a real-time view of the agent's effectiveness and tracks overall accuracy, identifies opportunities for improvement, and identifies solutions to help troubleshoot issues.

Model input

The models rely on two primary categories of input data:

  • Input Name and Supporting Information: name that led to the generation of potential exposure to a sanctioned entity, individual, or security. This also includes supporting information for the input name, such as the entity type, gender, address, date of birth, and location.
  • Watchlist Entity and Supporting Information: name, identifying data, entity type, and supporting information for the watchlist entity associated with each hit.

Model overview

The agent uses models designed to support consistent and explainable decisions:

  • Name Matcher model: provides accurate and efficient matching of names across different categories using specialized algorithms tailored to the unique characteristics of each name type.
  • Decision model: uses configurable rules, thresholds, and model outputs to produce a final disposition for the sanctions screening hit, leveraging the outputs of all features, along with predefined checks for specific scenarios.
  • High-level processing flow
  • Key agent components
  • Model input
  • Model overview

Was this page helpful?

Connect

Need help? Support

Want to learn? UiPath Academy

Have questions? UiPath Forum

Stay updated