AI Moments identifies meaningful segments in your video content and presents them as structured starting points for editing.
Instead of watching an entire source to find usable excerpts, you can review detected Moments, select relevant passages, and refine them in the new Clip Editor (V2).
A Moment can contain:
- IN and OUT points: the detected start and end of the segment.
- Moment type: what the segment represents.
- Title and descriptive information: generated information about the segment.
- Metadata: fields defined for that Moment type.
- Score: additional information about the detected result, where available.
AI Moments supports multimodal detection: depending on the configuration, analysis can use information from transcripts, visual content, on-screen text, and other signals. Not every signal is necessarily available or used for every source.
A Moment is not a finished clip. It identifies a segment worth reviewing. You decide whether to use it, how to cut it, and what to publish.
Categories, Types, and Moments
These three concepts serve different purposes:
| Concept | Purpose | Example |
|---|---|---|
| Moment category | Groups related Moment types and provides shared configuration context. | News |
| Moment type | Defines what to detect and which metadata to generate. | Weather / Forecast |
| Moment | An individual segment detected in a source. | A weather forecast between two timestamps |
The documented taxonomy includes categories such as News, Sport, and Talk / Debate, with specialized types and metadata within each category. The catalog available in your environment may differ.
Build your own custom Moment Categories and Types
AI Moments come with a set of pre-built moment categories and moment types. You can create your own categories and/or types to fit your exact business and audience.
To customize existing types, add new types in a category, or create new categories, check Managing AI Moment Types and Managing AI Moment Categories.
Limitations
Editorial review is required
AI Moments are here to help you save time for your live clipping sessions, but detection and generated metadata can be imperfect. Review the source rather than relying solely on a moment title, label, or score.
Results depend on content and configuration
Detection quality depends on the source, available signals, and how clearly the type is defined. There is no documented guarantee that every relevant passage will be detected.
Advanced automation is separate
AI Moments detection does not itself imply automatic highlight assembly or automatic clip creation, branding, and publication. Storybot and Moment-triggered workflows are separate capabilities.
Availability varies by deployment and permissions
Source workflows, configuration controls, and card actions depend on your deployed version and account access.
FAQ
What is the difference between a category and a type?
A category groups related types and provides shared context. A type defines a particular detection target and its metadata.
What is the difference between a Moment and a clip?
A Moment identifies a source segment. A clip is content you create and refine through the editing workflow. Detecting a Moment does not mean it has been published.