- Overview
- Financial Crime Compliance (FCC) solutions

Financial Services Solutions user guide
Frequently Asked Questions (FAQs)
Whether you want to learn more about the agents or have doubts on how to proceed with configuring your agents, browse the list of FAQs below.
Access and availability
What are the deployment options for the Financial Crime Compliance (FCC) solution? Is the FCC solution available on-premise?
The FCC solution is available only on UiPath Cloud. However, customers must be provisioned in Automation Cloud™, as license activation, entitlement management, and consumption tracking are handled there. Additional deployment options, including on-premise, are available directly from WorkFusion.
Is the FCC solution available in all regions?
The FCC solution is currently available in North America, followed by the United Kingdom, Austria, and New Zealand, Europe, Middle East, and South America. Additional regions will be planned for future releases.
What data residency is supported?
We can support any data residency supported by Microsoft Azure.
What languages are supported?
The models are primarily trained and fine-tuned on English data and other Latin-script languages. This means that other languages can be supported but the underlying models might need to be retrained to ensure optimal performance. Transaction Screening Alert Review and Name Screening Alert Review are typically performed in English, regardless of the region. Adverse Media Monitoring can support articles in English and Spanish, whilst all other languages can be transliterated. All models can be finetuned on a new language, if required.
The solution's UI is currently only available in English.
Is the FCC solution available to all Enterprise customers?
The FCC solution will only be available for customers with a required license. You will only be able to use the solution once fully deployed and configured by our implementation partner.
Is the FCC solution available in Automation Cloud™ Public Sector?
No. The FCC solution is not currently available in Automation Cloud™ Public Sector (FedRAMP).
Is the FCC solution available to UiPath Community users?
No. The FCC solution is not available to Community users.
Is the FCC solution available to UiPath partners?
FCC solution is currently unavailable to UiPath partners for testing, selling, or delivery. We are working with a selected partner on FCC implementations, with additional partners enabled in a later phase.
How long does it take to implement? Do I need to buy Professional Services?
Typically, it takes up to two or three months to deploy and configure without any major customizations. Professional Services is required on top of the licensing cost, as all production use cases are implemented by our delivery partner.
Do you offer a POC/POV for the FCC solution? What are the requirements?
We offer a limited-scope POV for qualified customers. Customers need to provide a historical export of their alerts or entities (covering at least a 30-day period), along with their dispositioning decisions, to compare the results and measure the impact. Live system integration is out of scope and is only possible for production use cases. POVs are typically completed within two to four weeks.
Data and security
Are the FCC agents deployed in my Automation Cloud instance?
No, FCC agents are not deployed in your Automation Cloud instance. It is a separate UiPath-managed Azure deployment, where each customer has their own instance (single tenant). Each FCC deployment is integrated with your Automation Cloud for licensing, monitoring, and telemetry purposes.
Is the solution fully deployed in the UiPath environment? Does any customer data leave UiPath Cloud?
Yes, the solution is fully hosted and managed by UiPath and data does not leave the UiPath environment. Data is stored in your own instance and is not shared with any other customer or third-party vendor. All data is encrypted in transit and at rest.
Are you using my data to train your models?
No. We only train your own model in your own instance, not the underlying parent model. Our approach is analyzing the patterns and generating synthetic data. The parent model is only trained with your data if you explicitly opt in and provide written permission.
Can I bring my own subscription?
Yes. If you choose to use the LLM features, you need to bring your own API key.
Is the FCC solution part of the AI Trust Layer (AITL)?
No. The FCC solution is not part of the UiPath AI Trust Layer, as it does not use any third-party LLMs managed directly by UiPath. The use of LLMs requires your own subscription.
Licensing and pricing
How do I buy an FCC solution? How much does it cost?
The FCC solution requires a separate license, which is available to both Unified and Flex customers. Evan, the Adverse Media Monitoring agent, is priced based on the volume of entities or searches, whereas Tara, the Transaction Screening Alert Review agent and Evelyn, the Name Screening Alert Review agent, are priced based on the volume of alerts.
For example, here is how each agent would be priced based on license type:
| License Type | Pooling Rule | Agent | Unit Consumption | Notes |
|---|---|---|---|---|
| Unified | FCC units can be pooled with Platform units. | Evan, the Adverse Media Monitoring agent | 1.2 FCC/Platform unit per entity | Applicable to both historical and ongoing searches |
| Tara, the Transaction Screening Alert Review agent | 0.6 FCC/Platform unit per alert | — | ||
| Evelyn, the Name Alert Review agent | 0.6 FCC/Platform unit per alert | — | ||
| Flex | FCC units cannot be pooled with agents/AI units. | Evan, the Adverse Media Monitoring agent | 6 FCC units per entity | Applicable to both historical and ongoing searches |
| Tara, the Transaction Screening Alert Review agent | 3 FCC units per alert | — | ||
| Evelyn, the Name Alert Review agent | 3 FCC units per alert | — |
Is there a minimum volume requirement to get started?
Yes. The minimum volume requirement is the following:
- Evan, the Adverse Media Monitoring agent: 60K entities (applicable to both historical and ongoing searches)
- Tara, the Transaction Screening Alert Review agent: 120K alerts
- Evelyn, the Name Screening Alert Review agent: 120K alerts
Do I need to buy all the agents?
No. The agents work independently, and you can choose to start with one or multiple agents.
Do you charge for dev or test consumption or only for production?
Yes, we charge for all types of consumption, regardless of the environment or operation. We recommend reserving an extra 10% of your annual volume for pre-production development and UAT.
Runtime
How do I ingest my data?
We offer various data ingestion options, depending on the specific agent:
Evan, the Adverse Media Monitoring agent
- As part of the Adverse Media Monitoring file ingestion BP, Evan looks for a new CSV file in the configured S3 directory every minute and initiates a new investigation if a new file appears. File ingestion allows users to upload a list of article URLs and entity information to be screened instead of having the agent search for and retrieve data from external providers.
Tara, the Transaction Alert Screening Review agent
- Pre-built screening system connectors via APIs
- Transaction Message format: FUF, SWIFT or SWIFT_MT, ISO20022, SWIFT_MX, SWIFT_NPP, SWIFT_ADK , or custom message
- Custom screening system connectors via APIs
Evelyn, the Name Alert Screening Review agent
- Connects to an FI’s name screening system and ingestion of alerts, hits, name input information, and watch list entity (WLE) list data. This connection is typically established through an API. The AI Agent may also ingest information from a Customer Lifecycle Management (CLM) system if demographic information on the input name is not stored in the screening system.
How do I consume the outputs?
- Tara, the Transaction Alert Screening Review agent, produces a structured output that contains decision, justification, and reason code information.
- Evan, the Adverse Media Monitoring agent, creates a detailed audit trail as either a standardized PDF (such as a narrative report) or HTML format. The report contains the analyst(s) decisions on article materiality, cited articles, a confidence score for each article, static URL, keyword citations, full article text, as well as the final decision and adjudication for the investigation. The core AMM BP completes with standardized output generation. This task creates REST API responses, prepares investigation results for downstream systems, and finalizes the investigation lifecycle.
- Evelyn, the Name Alert Screening Review agent, uses a combination of decisioning steps to either disposition hits and alerts as False Positives or Needs More Information. The final output for Evelyn includes hit disposition and each factor decision and supported information.
Can I reconfigure the agent if I'm not happy with the outputs?
All FCC agents have configuration settings and custom rules that you can adjust as needed. This will be an iterative process during the implementation phase to ensure optimal performance. If you wish to adjust any settings post-deployment, you need to raise a support ticket. Small changes can be covered by the UiPath support team, but major changes might require Professional Services.
Do you provide confidence scores? How are they calculated?
Yes, confidence and similarity scores are available for each underlying model, which can be used in configuration thresholds:
| Tara, the Transaction Screening Alert Review agent | Evelyn, the Name Screening Alert Review agent | Evan, the Adverse Media Monitoring agent |
|---|---|---|
| The final decision is determined by configurable rules. In the NER model each extracted entity is assigned a confidence score representing the model's certainty in the prediction (Token-Level Probabilities, Entity-Level Aggregation, and Confidence Thresholding). The Name Matcher model computes a similarity score of two names. | The final decision is determined by configurable rules. The Name Matcher model computes a similarity score of two names. | The Sentiment Adverse model produces confidence score based on a probability of likelihood The Focal model produces confidence score based on a decision tree consensus No confidence scores are generated for the LLM models as LLMs do not produce meaningful confidence scores |
Do you provide any Human-in-the-loop (HITL) capabilities?
Yes. HITL is a core capability of each agent workflow. For adverse media, users review prioritized news articles. These are the articles that are deemed most relevant and require human investigation. For name and payment alerts, users review hits that were not auto-adjudicated and require human investigation.
Can I use Action Center or Apps to review manual tasks?
No. All adverse media tasks are reviewed in a standalone User Interface (UI) built specifically for this agent. All name and payment alert tasks are reviewed directly in your screening tools. There is no new UI for these two agents.
Can the FCC agents learn and improve?
Evan, the Adverse Media Monitoring agent, and Tara, the Transaction Screening Alert Review agent, can be retrained and improved with customer data, but Evelyn, the Name Alert Review agent currently does not have a feedback loop and does not learn from additional data. Name Sanction Screening is based on different algorithms and models, not machine learning. Each model has its own distinct input and output specification.
Maintenance and monitoring
Can I monitor the performance of the FCC agents? What performance metrics are available?
Yes, each agent offers pre-built analytics and customizable dashboards with various performance metrics. Examples include but are not limited to: alert, article automation rate, error rate, true positives, false positives, missed true positives, missed false positives.
Can I track the consumption of FCC agents?
Yes, you can monitor the agents’ real-time consumption in your Automation Cloud Admin portal. Navigate to Licenses > Consumables to see your available and consumed FCC units (Flex) or FCC platform units (Unified).
How often are the FCC agents upgraded? What does it entail?
A new agent version is typically released once a month. All upgrades are managed by UiPath and do not require any actions on your side. You will be notified when a new version becomes available, and it will be up to you to determine if and when you want to upgrade. During the upgrade process, you may experience downtime of up to four hours. We will ensure a minimum impact on your operations.
What level of support is included? How do I raise a support ticket?
You can raise a support ticket directly in the Customer Portal. All tickets are triaged and resolved as per your UiPath support tier, coverage, and response SLAs.
- Access and availability
- What are the deployment options for the Financial Crime Compliance (FCC) solution? Is the FCC solution available on-premise?
- Is the FCC solution available in all regions?
- What data residency is supported?
- What languages are supported?
- Is the FCC solution available to all Enterprise customers?
- Is the FCC solution available in Automation Cloud™ Public Sector?
- Is the FCC solution available to UiPath Community users?
- Is the FCC solution available to UiPath partners?
- How long does it take to implement? Do I need to buy Professional Services?
- Do you offer a POC/POV for the FCC solution? What are the requirements?
- Data and security
- Are the FCC agents deployed in my Automation Cloud instance?
- Is the solution fully deployed in the UiPath environment? Does any customer data leave UiPath Cloud?
- Are you using my data to train your models?
- Can I bring my own subscription?
- Is the FCC solution part of the AI Trust Layer (AITL)?
- Licensing and pricing
- How do I buy an FCC solution? How much does it cost?
- Is there a minimum volume requirement to get started?
- Do I need to buy all the agents?
- Do you charge for dev or test consumption or only for production?
- Runtime
- How do I ingest my data?
- How do I consume the outputs?
- Can I reconfigure the agent if I'm not happy with the outputs?
- Do you provide confidence scores? How are they calculated?
- Do you provide any Human-in-the-loop (HITL) capabilities?
- Can I use Action Center or Apps to review manual tasks?
- Can the FCC agents learn and improve?
- Maintenance and monitoring
- Can I monitor the performance of the FCC agents? What performance metrics are available?
- Can I track the consumption of FCC agents?
- How often are the FCC agents upgraded? What does it entail?
- What level of support is included? How do I raise a support ticket?