Add Multimedia Transcript Summaries
The Add Multimedia Transcript Summaries opportunity identifies pages where important information lives in video or other media without transcripts or short text summaries that AI agents can read. It introduces AI-generated transcript summaries grounded in the media and surrounding page context. It helps recover key brand information that would otherwise be missed by making multimedia content understandable to AI agents.
For each affected URL, you can review the proposed Content Patch, Implementation details (for example, the target CSS selector and operation) and Rationale, then deploy with Optimize at Edge so agentic traffic receives the enriched HTML with no Content Management system (CMS) changes required.
How it fixes the problem
Fixes are applied using Optimize at Edge, which:
- Serves a pre-rendered HTML snapshot to AI agents.
- Enriches the page with transcript summary text in the HTML they retrieve (for example, near the relevant inline video).
- Works at the CDN layer (no CMS changes).
- Is AI-only — no impact on human visitors or SEO bots.
- Deploys in minutes and is fully reversible from the LLM Optimizer interface.
How it works
LLM Optimizer flags high-traffic pages where machine-readable text is missing for embedded media, based on your configuration and page structure. Affected URLs appear in the URLs with suggestions table on the Current Suggestions tab, where you can expand a row to inspect each Content Patch, how it will be applied, and why it is recommended.
For each page, you have:
Multimedia Summary – Structured summaries derived from video content.
Preview – Before and after page comparison.
The URLs with suggestions table lists pages where transcript or summary text would help agentic discovery. Suggestions are organized into Current Suggestions, Fixed Suggestions, and Ignored Suggestions. For each URL you can:
- Expand the row to view the Content Patch text, Implementation details (including the planned DOM operation and CSS selector), and Rationale for the change.
- Preview the before and after comparison for agentic traffic.
- Mark as Fixed if you addressed the opportunity outside LLM Optimizer.
- Ignore suggestions that are not relevant.
You can edit patch text from the row when supported (pencil control), then use the row checkboxes to select what to deploy. The footer shows how many are selected and provides Mark as Fixed, Ignore Suggestions, and Deploy optimizations.
Deploying the optimization
When you are ready to publish at the edge, click Deploy optimizations. A Deploy to Edge dialog lists the URLs, selectors and operations you are about to run. Review the list, then choose Deploy or Cancel.
After a successful deploy, Deployment Complete confirms how many optimizations went live. Close the dialog and open Fixed Suggestions to verify status.
Fixed Suggestions and View Live
On Fixed Suggestions, deployed URLs show Optimized in the status column. Expand a row to review the live Content Patch, Implementation details and Rationale . Additionally, you can use Details for analytics or View Live where available to confirm what agentic traffic receives.
To roll back in bulk, select the optimized rows using the checkboxes, then use Rollback in the header.
Rollback
If you change your mind, you can roll back a deployed optimization. From Fixed Suggestions, select the rows to revert, then click Rollback.
The Rollback dialog lists the suggestions that will be rolled back and warns that deployed optimizations will be removed from the live path for agentic traffic. Confirm, then click Rollback or Cancel.
When the operation finishes, a Successfully Rolled Back summary appears; close it to return to the dashboard.
Try it in the demo
Explore the Add Multimedia Transcript Summaries workflow in the Frescopa demo.