Source: https://vaultcrmhelp.veeva.com/doc/Content/CRM_topics/VeevaAI/Overview.htm

## Vault AI in CRM Overview

Vault AI brings agentic AI capabilities into the core of Vault CRM, providing intelligent agents to help users find information, prepare for customer engagements, and complete tasks more efficiently. Vault AI is embedded throughout users’ natural workflows, providing easy access to context-aware AI features including interactive chat and search, pre-call planning, and voice entry.

Vault AI connects a Large Language Model (LLM) with a specialized Vault CRM data layer, ensuring our AI tools are grounded in accurate, customer-specific data. The data layer securely accesses, formats, and structures information from your CRM instance, giving the LLM the necessary context to provide precise responses tailored to your organization and users.

Agents are at the core of Vault AI. Each agent is designed to perform a specific function— for example, reviewing free text, guiding call planning, or searching for media — and works in coordination with others through AI Chat. This agent-based design allows Vault AI to deliver targeted, contextually relevant assistance across the CRM experience while maintaining accuracy, security, and consistency with your organization’s data.

Vault AI agents use Anthropic's Claude as the LLM, hosted on AWS Bedrock.

A Veeva AI license is required to use this functionality.

Not all features are available on all platforms. For a list of available platforms for each feature, see [Platform Functionality](../Platforms.htm).

[![Closed](../../../Skins/Default/Stylesheets/Images/transparent.gif)Users and Roles](#)

Vault AI in CRM allows users to perform the following actions:

#### Business Admin Users

Access Agent Monitoring information:

* Monitor AI usage — Monitor user adoption, analyze AI performance, and understand usage trends. With this data, organizations can make informed decisions to improve the user experience, ensure accuracy, and demonstrate the return on AI investment. For details, see [Monitoring Agents, Agent Actions, & LLM Token Usage](https://platform.veevavault.help/en/lr/1000971/).

  The Agent Instance and Agent Action Execution tables update every 6 hours for Vault CRM standard agents. These tables update immediately for custom agents. Tracing Agent Actions is not available for Vault CRM standard agents.

#### End Users

Access agentic AI functionality to complete tasks:

* Agentic Call Report — Generate call reports from dictated notes and check for compliance
* Agentic Email — Assist in composing Approved Emails
* Agentic Insight Quality — Check the quality and completeness of free text entries, as they are captured
* Agentic Account Summary — Prepare for customer engagements by receiving AI-driven context and suggested actions
* Agentic Media — Quickly find and summarize relevant media library content using semantic search
* Agentic Text Monitoring — Enter free text notes that are automatically checked for compliance issues
* Agentic Voice — Capture information by dictating notes that are intelligently mapped to Vault CRM records and fields

#### Compliance Teams

Leverage the following agentic AI functionality:

* Agentic Text Monitoring for Home Office — Retrospectively analyzes end users' text inputs for potential compliance concerns

### Related Topics

[Configuring Vault AI in CRM](Configuration.htm)

[Using AI Chat](UsingChatUI.htm)

[What is Vault AI in CRM?](https://vaultcrmflightpath.veeva.com/doc/Content/VeevaAI/AIAgents/1.Intro.htm)