AI-driven tagging automates strategic metadata assignment in e-Detailers, portfolio e-Detailers, emails, email fragments, AEM websites of the Commercial content type, and external light presentations created from DOC, DOCX, PDF, PPT, or PPTX files. Powered by models trained on your business data, this intelligent system analyzes content to apply relevant tags automatically.
With the AI tagging agent, you can:
Streamline tag assignment. AI-powered analysis accelerates the process and reduces manual effort in metadata management.
Ensure tagging accuracy. Maintain consistent tag assignment across content item types while adhering to your established taxonomy.
Optimize the tagging workflow. Tag content properly according to your business requirements and strategic guidelines.
Enhance user experience. Automatically assign relevant tags tailored to your content, reducing time spent on manual work.
Refine suggested tags. Validate and adjust automatically assigned tags to align with specific business requirements, markets, and campaigns.
The auto-tagging feature is available for users with respective permissions. Contact the administrator for details.
To ensure AI-driven tagging:
For AEM websites, assign product metadata to placeholders in Veeva Vault PromoMats. The combination of brand, indication, and subtherapeutic area values, along with related key messages, must be mapped to that product in eWizard
Centralized tagging.
e-Detailers and portfolio e-Detailers
Auto-tagging is triggered when you click PUBLISH to send the e-Detailer or portfolio e-Detailer to Veeva Vault PromoMats as a PDF for approval.
To launch AI-powered tag generation for your content item:
Open it in eWizard Centralized tagging.
Click the
AI SUGGESTIONbutton.
The AI-powered tagging starts automatically.
You're redirected to the auto-tagging progress page with the progress bar at the bottom. Click the NOTIFY ME WHEN IT IS DONE button in the bottom-right corner to receive an email once the AI agent finishes generating tags.
The user who initiated auto-tagging receives an email notification containing a link to the tagged content item.
If AI-powered tag generation fails, the starting Centralized tagging page opens with a warning that structural elements may require manual tagging. Tag the content item manually or restart auto-tagging.
Contact the administrator to configure the following:
Email notifications when the auto-tagging completes.
Email notifications with an attached CSV report for specified users when AI-powered tagging fails. The report contains details on each structural element where the AI agent couldn't apply tags.
Progress page behavior for failed auto-tagging.
Once auto-tagging is complete, the AI tagging agent populates the fields and selects Key message not required checkboxes. Up to 3 Global key message/Local key message tags can be assigned to each structural element. The 
symbol marks fields and checkboxes handled automatically, with fields highlighted by blue frames.
For the Global key message/Local key message field, the AI tagging agent can offer several alternatives with different confidence scores:
— High
— Medium
— Low
These alternatives are placed above the key message values that match the set Brand/Indication/Subtherapeutic area values.
Hover over any confidence score marker to view the reasoning behind the AI-suggested key message.
Contact the administrator to configure automatic key message selection based on required confidence scores.
If you delete AI-assigned key messages, a rejection popup appears.
In the popup, do the following:
Choose a reason for rejecting the key messages:
If you delete a single key message with the
button—select a reason only for that specific message.If you remove key messages by clearing their checkboxes—select a reason for each rejected key message.
2. Click SAVE to confirm your choices.
After the content item is published to the target system, all saved rejection reasons are automatically sent to the repository.
The rejection popup stays open until you select and save a reason or reload the tagging page.
Contact the administrator to configure the list of rejection reasons.
The AI tagging agent can also automatically select the Key message not required checkbox in the following cases:
When the AI-generated key messages have a low confidence level.
Contact the administrator to configure automatic selection of the
Key message not requiredcheckbox based on the required confidence level.
When the AI-powered tagging agent detects that key messages aren't necessary for the content analyzed.
e-Detailers, portfolio e-Detailers, emails, email fragments, and external light presentations created from DOC, DOCX, PDF, PPT, or PPTX files
If a content item was auto-tagged, AI-powered tag generation can't be run again.
You can rerun the process in the following cases:
Some structural elements were not auto-tagged. The AI tagging agent processes only those elements.
Structural elements were added or modified. The AI tagging agent processes only the newly added or modified elements.
Brand, country, or language metadata was changed. The AI tagging agent processes the entire content item.
Version of the source VVPM document was changed for external light presentations. The AI tagging agent processes only the newly added or modified elements.
AEM websites
The AI tagging agent processes AEM website pages as follows:
If a page doesn't have tag containers, the agent assigns key messages to the page.
If a page has tag containers, the agent automatically selects the
Key message not requiredcheckbox.
If the AEM website was auto-tagged, you must wait 24 hours before running AI-powered tag generation again.
However, you can rerun it within 24 hours if any structural elements were not auto-tagged. In this case, the AI tagging agent processes only the untagged elements.
AI-suggested tags require human validation for accuracy.
If you've manually tagged your content item before launching auto-tagging, the Brand, Indication, and Subtherapeutic area tags are preserved and not overwritten.
External light presentations created from DOC, DOCX, PDF, PPT, or PPTX files
A manually entered value in the Name field is always overwritten by the AI-suggested variant.
The 
symbol marks manually assigned Brand, Indication, and Subtherapeutic area tags, as well as the Name value in external light presentations created from DOC, DOCX, PDF, PPT, or PPTX files when these values match the AI-suggested ones.
As for key messages, the system applies AI-suggested variants based on the number of existing manual key messages:
AI suggests 3 key messages: The system replaces all manual key messages with the 3 AI-suggested ones.
AI suggests 2 key messages and 1 manual exists: The system keeps all 3 key messages.
AI suggests 2 key messages and several manual exist: The system adds the 2 AI-suggested messages, keeps the first manual message, and removes the rest.
AI suggests 1 key message and 1 manual exists: The system keeps both key messages.
AI suggests 1 key message and several manual exist: The system adds the AI-suggested message, keeps the first 2 manual messages, and removes the rest.
After auto-tagging, manually add key messages if required. The system keeps both the AI-suggested and manually added key messages.
You can change autopopulated tags by manually selecting different values and adding new tagging fields, if required. When a field is edited manually, the 
symbol and blue highlighting disappear.
To finalize the auto-tagging and send the AI-generated tags to the repository, click SAVE.
Additional scenarios
Reselecting a previously rejected key message for the same structural element
If you save a reason for rejecting an AI-assigned key message and then manually reselect the same key message for the same structural element, the reselected key message is sent to the repository after the content item publication to the target system. If you delete the key message again after reselecting it, the rejection popup doesn't appear, and the previously saved rejection reason is sent to the repository.