June 25, 2026

New features

• AI-suggested banner layouts in eWizard Editor

  Use the eVa AI chat to generate ready-to-use banner layouts. The Copywriter agent populates a predesigned template with content from your knowledge base according to language, brand, and country metadata selected when creating a banner. You can preview and apply variations to the selected banner size. The applied variation is then available for further modification.

  eVa AI reduces manual layout creation effort and automates content drafting, helping you produce compliant banners faster. Configuration required

Improvements

• eVa AI now supports searching for texts, claims, images, and documents in the client-specific knowledge base using exact or partial Veeva Vault document numbers, document names, or claim numbers. All other metadata used as search criteria requires exact values. This improves search efficiency by allowing faster retrieval of relevant assets with less strict metadata input requirements.

• MLR Acceleration

  • The MLR Acceleration agent now validates references more intelligently by analyzing both the content and the supporting sources. The system automatically identifies text elements that require citations, verifies whether attached references are relevant to the claims made, and flags content with missing or insufficient supporting references.

    This improves citation accuracy, strengthens compliance, and helps ensure claims are supported by appropriate evidence.

  • Applying suggestions from the RESULTS panel now provides immediate visual feedback. The APPLY button displays a loading state and is temporarily blocked to prevent duplicate actions.

  • The MLR Acceleration agent workflow is improved by keeping validation behavior consistent across all content edits. Validation results are preserved regardless of whether changes come from MLR Acceleration agent recommendations or manual user updates.

    If validation results belong to an earlier content version, the system prompts revalidation to display results for the current version.

  • Validation checks now run in a configurable order, ensuring that the most critical checks are performed and displayed first. This makes it easier to identify and resolve the most important issues. Configuration required

  • Each pharmaceutical marketing code document is matched to content using configurable metadata criteria, such as brand, country, and language, reducing irrelevant validation results and making compliance reviews more accurate and efficient. Configuration required

Admin setup

• To set up banner layout generation for the eVa AI chat, update the eva object in the environment configuration file as follows:

  • Add the banner_variations string value to the editSet array in agents → supervisor → agentsToManage, and the banner_variations object in the agents → supervisor to activate the feature.

  • Adjust the bannerLayout object in modules to configure text and image search in the client-specific knowledge base.

• To define the sequence in which validation checks are executed and control the order in which results are displayed, add the executionOrder field to any validation rule in the configuration. This allows prioritization of critical validations.

• To control how pharma codes are filtered and matched during content validation:

  • Add the filterConfig object under pharmaCodesConfig to define which content metadata fields are passed to the pharma code filtering logic.

  • Add a named object for each pharma code document to define the acceptable metadata values for that document and the pages included in validation.