Use of customer data with IXP

Purposes

We process the information used with our services for the following purposes:

• Provide our services
• Machine learning (ML) and product enhancement and development

Provide our services

In Communications Mining™, we use messages provided at design time to train customer-specific models, and while running historical analytics and automations in your environment. This can include passing message data through activities for human-in-the-loop validation in other components of the Platform, such as Apps and Action Center. The model is trained for the specific customer, and the customer has control over the data input. For more details on how Communications Mining works, check the Communications Mining overview page.

In Generative Extraction for Unstructured and complex documents, we use documents provided at design time to validate performance of predictions for the extraction models, and at run time to generate predictions in your environment. This can include passing document data through activities for human-in-the-loop validation in other components of the Platform, such as Apps and Action Center. You control the configuration of the project used to request predictions and have control over the data input.

We process the messages, documents, and configurations provided at design time and at run time to ensure the IXP product and Platform meet quality standards. This means that we monitor, test, evaluate, and measure performance across all functionalities and infrastructure, to ensure high performance of the model for your use cases.

We process telemetry data, such as usage data, services latency, traffic peaks, as well as user and account identifiers (i.e. account IDs, tenant IDs, or user e-mail addresses). The purpose is to make sure that our platforms operate smoothly, and to optimize user experience based on the feedback and usage patterns we observe. We use data analytics to measure and evaluate the performance of our services.

We process account information, such as username, password, and so on, to provide you with access to the services and to secure your account.

We may process the data you provide for support or, as necessary, for troubleshooting purposes, in accordance with our Support terms.

Machine Learning (ML) and product enhancement and development

We use messages sent to the UiPath platform at design time to train the UiPath Communications Mining model, enhance, and improve the product. This way, we can provide you with a simpler, faster, and more intuitive user experience, along with a higher automation performance at run time. For example, when we improve the performance of the pre-labeling capability at design-time, we reduce the manual work you need to do to build high-performing, customized ML models to process messages. This helps reduce the time and effort required for building enterprise-ready automations.

Note:

We do not use data for machine learning if you have a BAA in place, if your contract forbids it, or if you have submitted a Support ticket requesting to opt out of Data Usage for your account.

Data storage

Data is stored when using this product as selected by you and in accordance with our data residency page, available at Data residency.

Integration with Generative AI services

The Communications Mining capability has some features which rely on integrations with third-party Generative AI services, and the Unstructured and complex documents capability cannot be used without these services. All such integrations go through the UiPath Trust Layer. This component ensures that all integrations comply with the same high standards of security and compliance, and that no data is persisted by any third parties for any reason.

Safeguards

We employ the following safeguards to achieve the previously-mentioned purposes:

• Customers who have a BAA in place are excluded from Data Usage for product improvements.
• Customer data is deleted at contract termination or upon request.
• Customer data is encrypted. Access is strictly limited, and any access is logged and audited.
• Training data remains in the hosting region selected by the customer.
• Both run-time data and training data are encrypted.
• The customer can delete, modify, or export design-time and run-time data. The customer can request purging of data.

Opt-out mechanism

We have an opt-out mechanism for model training.

To opt out, apply the following steps:

1. Submit a Support ticket in which you can request to opt out.
2. Ensure that the contractual agreements reflect your choice regarding data usage for training purposes.