Using the standard (improved video + element definitions) finalized through STEP5 (selection → extraction → classification → improvement → verification), this feature uses AI to view and align the repeated work videos of target persons (junior staff, new hires, etc.) and visualizes cycle time comparison, delay per element, and AI coaching.

This makes it immediately clear who is taking time or showing variation in which elements, and clarifies the priorities for training and on-the-job training.

This article is STEP6 of the work analysis series. For the full picture, see Work Analysis overview and getting started.

Pre-setup: confirm the standard

Before running a comparison, you need to progress through STEP1–5 (selection through verification) to finalize the standard video and element definitions. The comparison step cannot be executed without a confirmed standard.

Add comparison (choose target person and video)

1. Open "Compare" (STEP6) on the left rail
2. Click [Add] in the top right
3. Set the following in the dialog that appears
   • Target person: select from team members (to record whose work video it is)
   • Work video: select the target person's repeated work video from the asset library
4. Click [Start] to begin video analysis with AI (you will be notified by email when complete)

Comparison list and details

Added comparisons appear in the target person list. You can confirm the target person, video, analysis date, and status. Clicking a row opens the detailed screen (comparison layout).

• During analysis "Analyzing" is displayed, "Partially completed" for partial failure, and "Failed" for failure
• Completed comparisons show no status display (only highlighted when attention is needed)

Comparison layout (view standard and target person side by side)

In the detailed screen, the standard video appears on the left and the target person's video on the right, with playback position synchronized between the two. In the comparison timeline at the bottom, the target person's delay is visualized with red shading.

Cycle time (stacked bar graph)

• Left end = standard cycle time, followed by each cycle for the target person
• Each bar displays elements stacked together = cycle time
• Standard (blue dashed line) and target person average (orange dashed line) reference lines are shown
• Click a cycle to switch to detailed comparison for that cycle

Delay per element

Shows the difference when the standard is set to 0, displayed as a line graph by element.

• Selected cycle: solid line (orange) + red shading
• Average: dotted line (light color)

Variance per element (with 2 or more cycles)

For each element, you can compare the target person's shortest – longest range, average, selected cycle, and standard all in one screen. You can immediately see whether performance is consistent or shows wide variation.

AI coaching (improvement suggestion)

From the [Improvement suggestion] tab in the right panel, AI can generate personalized coaching for this target person.

1. Click [Generate suggestion] in the top right
2. AI proposes improvement points at the element level based on differences between the standard and target person
3. Each suggestion displays priority (high / medium / low), the target element, point, and advice

If you're not satisfied, you can try again with [Regenerate suggestion].

Change target person / adjust cycle

• You can also change the target person later from the top of the detailed screen (if you need to correct whose data it is)
• From [Cycle adjustment] in the top right, you can fine-tune if the AI's automatic cycle boundaries are offset for the target person
• If you're not satisfied with the analysis results, you can try again with [Re-analyze]

Handling failures and partial failures

• Failed: "AI observation" and "Error details" are displayed. Confirm the cause and execute [Re-analyze]
• Partially completed: If there are failed sections, you can use [Re-analyze failed chunks only] to re-run just those parts (does not consume AI count)

Tips for best use

• Refine the standard before comparing: If the standard's element definitions are rough, comparison results will be rough too. The more carefully you create the model in STEP1–5, the higher the comparison accuracy
• Prepare multiple cycle videos: When you prepare 3–5 or more cycles of the target person's repeated work video per analysis, variation and average are calculated more accurately
• Use as on-the-job training material: Concrete statements like "You're delayed by X seconds in element ◯◯, you'll speed up by moving like this" greatly change conversations on the shop floor
• Re-run comparisons periodically: As a measure of training effectiveness, re-running comparisons of the same task monthly or quarterly lets you visualize improvement

Related

• Work Analysis overview and getting started
• STEP5 Verify: compare before and after improvement and export video