This explains how to run AI analysis later on materials where you've only uploaded the video without running AI analysis. Use this when you want to "register just the video for now and run AI analysis in batch later" or "I uploaded without narration, but I want AI to analyze it after all."
For details on what each option changes and how to choose, see AI processing options: items and how to choose them.
Videos that already ran AI analysis during upload are not covered by this article. If processing has failed, see How to re-run failed AI processing. If you want to redo a successful result, see the "Note: When re-running after successful processing" section in that same article.
Steps to execute
- Open [Materials List] in the sidebar
- Click on a video that hasn't been AI analyzed yet to open the material's detail modal
-
Click the [Run AI processing] button displayed at the bottom of the detail modal
- The AI processing options modal opens. Adjust the following as needed:
- Videos without narration (analyzed only by video): Check this for videos with no narration, videos with only background music or mechanical sounds, or videos with unclear speech
- Recognition Language (within [Advanced settings]): The language of the speaker in the video. Matching this improves transcription accuracy
- Video description: Entering a description of the video improves analysis accuracy
- Split granularity: How finely to split steps (Process / Step / Action / Single). "Step" is recommended when publishing as an SOP
- Target step count (within [Advanced settings]): Specify approximately how many steps to split the entire video into, or specify seconds per step. With the default [Auto], AI decides based on split granularity
- Reference procedure to use (within [Advanced settings] · β): Specify an existing SOP file and split according to the breaks in that material (see Specify a reference SOP in video analysis (β) for details)
- Text style: The sentence ending of step text (Imperative / Honorific / Suggestive). If a Procedure Authoring Policy is registered, you can also choose [Follow policy] (default)
- Subtitle style: The sentence ending of action subtitles in the video. You can specify this separately from step text (default is [Imperative]). To match the step text, choose [Match text]
- Click [Run], and the status changes to "Waiting." AI analysis will start in sequence
Analysis completes in roughly the same time as the video length. For longer videos, analysis continues even if you close the app, and you'll receive an email notification when it completes.
Launching from the AI status tag
You can also click the AI status tag in the top right of a material card (like "AI not run" for unanalyzed videos) to reach the same detail modal and AI processing options. This is convenient when you want to add multiple videos to the analysis queue in sequence from the Materials List.
Frequently asked questions
Q. Can I run analysis on multiple videos at once?
A. You'll press [Run AI processing] → [Run] one at a time. After execution, they're queued in sequence, so you can operate in batch by simply opening the next video's detail modal and pressing the button.
Q. What if I want to adjust the AI processing results later?
A. Once analysis completes, you can change options and re-analyze from [Check AI analysis results] → [Re-run AI] in the detail modal (processing overwrites the previous result). See the note section in How to re-run failed AI processing.
Q. What if I want to skip AI analysis during Step1 upload?
A. Cancel without pressing [Run] on the AI processing execution screen during upload, and the video will be registered without AI analysis. You can analyze it later using the steps in this article.