Deep AI that’s flexible, secure, and compliant for regulated companies of all sizes.
Cross-platform agentic AI, plus application-specific agents for deeper work.
Platform Agents operate across the Veeva Vault Platform, using the unified data and content within to perform general tasks like translation and analysis.
Application Agents operate inside Quality applications, performing deeper tasks using more specific context, prompts and safeguards.
We upgrade Vault AI Agents and release new ones three times a year. Customers can also configure and extend Vault AI Agents, and create their own Custom Agents .
Vault AI Agents use large language models (LLMs) from Anthropic and Amazon, hosted on Amazon Bedrock. Custom Agents use Veeva-hosted models or customer-provided models, hosted on Amazon Bedrock or Microsoft Azure AI Foundry.
Agentic AI built into Vault Platform
Direct, secure content and data access
Configure Vault AI Agents or create custom
Vault AI targets manual, time-consuming workflows to maximize productivity and free users to focus on higher-value work.
Vault AI supports mandatory human escalation points for operations that exceed defined risk thresholds.
Extend Vault AI agents, or build your own Custom Agents for unique use cases.
Access continuously improving Vault AI capabilities with three upgrades per year.
Ask questions about documents; summarize, translate
What are the key takeaways from this SOP?
Do the changes in this document require a training update?
Ask questions about objects; summarize, translate
What was the cycle time for this change control?
Summarize the key decisions to be verified from this Training Requirement Impact Assessment.
Understand changes between document versions
What has changed since the last version of this SOP?
Tell me more about this specific change.
Ask questions about data (objects) and document metadata
Do I have any upcoming periodic reviews?
Have there been any non-conformances for product X in the last 90 days?
Ask questions about your data (documents) in Vault
Find me cleaning SOPs for the Mexico facility.
What SOPs contain information about job roles?
Get assistance on SOPs and help on next steps
What are the cleaning steps for production line 4 in the Mexico facility?
Help me with root cause analysis for this non-conformance.
Veeva offers a high degree of granularity in how AI Agents can be enabled for users. Permissions can be set for individual Agents by User type or by individual Users, which cascades to permissions sets. External Users can use Vault AI within the parts of the Vault platform they have access to.
No. Veeva understands that your data and your documents are your intellectual property. We do not train models nor do we allow 3rd party subprocessors to train models on your data. We only allow inference — the process or reasoning on data — to happen with your data.
Veeva’s investment into compliance is unique in the industry. The Veeva compliance shared responsibility model defines the division of regulatory, validation, and security duties between Veeva Systems and the customer. Veeva provides the secure, validated cloud infrastructure and application capabilities, while the customer remains accountable for internal data integrity, standard operating procedures (SOPs), and user access controls. This is all true with AI as well.
Generative AI is inherently probabilistic and can make mistakes. Therefore, Veeva cannot guarantee 100% absolute accuracy for every single output or summary. We have done significant work in ensuring the LLMs have the right data to provide valid insights and helpful recommendations. This starts with our system prompt. This prompt sets the boundaries for the LLM. Next comes the individual agent’s prompts, data, and tools. The standard agents provided by Veeva as part of Vault AI go through rigorous evaluations designed to ensure the agent is effective. We also evaluate responses to ensure that they are appropriate.
Veeva QualityOne helps manufacturers deliver high-quality products that consumers trust. The cloud-based platform unifies applications, processes, and partners across content management, Quality Management System (QMS), food safety management, and training.
