Agentic Text Monitoring Overview
Agentic Text Monitoring uses AI to proactively analyze information entered in free text fields and flag potential compliance concerns before the information is saved. This intelligent safeguard enhances users’ ability to efficiently and compliantly plan, execute, and document interactions with HCPs, and it enables users to confidently add more detailed, valuable notes about HCP interactions. Organizations can then leverage these detailed notes to develop deeper business insights.
Agentic Text Monitoring considers the following for compliance:
- Inappropriate Content Screening - Restricted words, profanity, and inappropriate content, as defined in the LLM’s content filters
- Data Protection - Entry of protected health information (PHI) or personally identifiable information (PII) in unencrypted fields
- Promotional Compliance - Off-label product promotion. This applies to company products only.
- Product Restrictions - Mentions of products that are either restricted or not allowed for the selected account, location, and/or attendee accounts at the time the compliance check occurs
- Safety Vigilance - Reporting of adverse events, serious adverse events, or product quality complaints outside of designated safety systems
User inputs are checked in any language supported by the base Large Language Model (LLM). Agentic Text Monitoring’s responses are displayed in the language defined for the user in Vault CRM, as are Vault CRM actions, alerts, errors, and labels.
Agentic Text Monitoring highlights potential issues based on the above concerns and any supplementary reference documents or compliance-type documents provided, while respecting legitimate use cases and regulatory safe harbors. The country and state associated with the call address are considered when determining whether users’ text poses a concern, as are the user type, country, division, department, and title. This includes international context and local regulations, when applicable.
Customers can provide additional information to add a company-specific definition of compliance. This includes compliance policies, for example, standard operating procedures, work instructions, handbooks, and more. For more information, see Providing Compliance Documents.
Agentic Text Monitoring’s findings can be captured on Monitored Text records in Vault CRM, providing compliance teams with direct visibility to potential issues. For more information, see Capturing Agentic Text Monitoring Findings.
Not all features are available on all platforms. For a list of available platforms for each feature, see Platform Functionality.
Accessing Agentic Text Monitoring
To access Agentic Text Monitoring, select a monitored text field. The following entry point is supported:
- Agentic Call Report™ - Call Text field
Token Estimates
Token consumption may vary based on a number of factors, including the length and complexity of the user’s text, the number of relevant products, and setup choices like providing corporate compliance documents.
Based on testing during the product development process and early adopter usage, a reasonable estimate may be 15,000-25,000 tokens per run. Agentic Text Monitoring runs when text is entered in applicable fields, or when users save or submit the call report.
