Configuring Veeva AI for Vault CRM
Veeva AI in Vault CRM must be configured by an admin before it becomes available to users. The configuration process includes both the initial system setup (detailed below) and agent-specific setup to enable individual AI features.
Veeva AI in Vault CRM incorporates the following components.
Vault CRM Data Layer
Vault CRM’s data layer securely provides the Large Language Model (LLM) with contextual data from your CRM instance. This layer includes the Vector Database, which stores approved content as vectors. This ensures the model's responses are grounded in customer-specific data that the end-user already has access to view.
Large Language Model
The Large Language Model analyzes language, context, and intent of user prompts and input. It processes each request, generating responses grounded in the information retrieved from the Vault CRM Data Layer.
AI Settings
AI Settings provide the ability to define specific AI application behavior for a Vault or an application profile. For more information, see Configuring AI Settings.
AI Agents
- Pre-Call Agent — The Pre-Call Agent uses AI to intelligently assist users in planning and preparing for interactions with HCPs
- Voice Agent — The Voice Agent uses AI to convert your words into text, then intelligently maps that text to fields and records in Vault CRM
- Free Text Agent — The Free Text Agent uses AI to proactively analyze information entered in free text fields and flag potential compliance concerns before the information is saved
- Media Agent — The Media Agent uses AI to help users quickly find information about products and access relevant media for HCP interactions
Configuring AI Settings for
To simplify administration, Veeva AI for Vault CRM uses configuration records called AI Settings, stored on the ai_settings__v object. These settings define how AI features behave across your Vault and can be applied at either the global (Vault-wide) level or the profile level.
How AI Settings work:
- Global AI Settings control the default behavior for the entire Vault
- Profile-level AI Settings allow administrators to override global defaults for specific application profiles (for example, to tailor AI access or behaviors for Field Users vs. Managers)
- When both are configured, profile-level settings take precedence over global settings
- A Global AI Settings record is provided by default
To view or edit AI Settings:
- Navigate to Business Admin > Objects > AI Settings.
- Select one of the following:
- The Global AI Settings record for Vault level changes
- A profile-level AI Settings record
- Select Edit to make changes for one or more AI Settings.
- Select Save.
To create a profile-level AI Settings record:
- Navigate to Business Admin > Objects > AI Settings.
- Select Create.
- Enter a Name.
- Select an Application Profile to which the AI Settings should apply.
- Set the status to Active.
For additional information on creating profile-specific settings, see Configuring VMOCs or Settings for an Application Profile.
Configuring Veeva AI in Vault CRM for
To configure Veeva AI in Vault CRM:
- Navigate to Admin > Users & Groups > Permission Sets.
- Select the appropriate Permission Set.
- Select the Objects tab.
- Grant the following permissions:
Object Object Permission Object Types Fields Field Permission ai_settings__v R None - application_profile__v
- created_by__v
- created_date__v
- initial_salutation__v
- modified_by__v
- modified_date__v
Read -
Select the Agents tab.
-
Grant Execute access to the appropriate AI agents:
- Free Text Agent (text_monitoring_agent__v)
- Media Agent (media_agent__v)
- Pre-Call Agent (pre_call_agent__v)
- Voice Agent (voice_agent__v)
The agents listed above may require access to the Veeva AI Chat permission. See Using AI Chat for more information.
Veeva AI in Vault CRM functionality is not available until specific features are configured.
- Navigate to Admin > Settings > Veeva AI Settings.
-
Select Veeva AI in the Default LLM Connection field.
- Navigate to Business Admin > Objects > VMobile Object Configurations.
- Activate the VMOC for the ai_settings__v object for the iPad platform.

