What a report looks like.

This is a sample for illustration. Run TurboProc Scope on your own line to get a report with your real times.

AI Optimization Report

Device packaging, manual line

Per-step timing

Median across 5 runs. The step carrying the most recoverable time is flagged.

#	Step	Median	Best run	Variation
1	Place box on bench	4.1 s	3.8 s	0.6 s
2	Open box	5.3 s	4.9 s	0.9 s
3	Place device in box	7.0 s	6.6 s	1.1 s
4	Add packaging foam Bottleneck	12.8 s	7.9 s	4.9 s
5	Close box	6.2 s	5.7 s	0.8 s
6	Remove box to pallet	6.2 s	5.9 s	0.5 s

Where the time is going

Step 4 (Add packaging foam) is the bottleneck: it is both the longest step and the most variable, swinging 4.9 s between the best and median run. The other five steps are stable to within ~1 s. The AI's read: foam is being retrieved and oriented at the moment of use rather than pre-staged, so the operator's time-to-place depends on how the foam happens to be sitting in the bin. That is a staging problem, not an operator-speed problem.

Recommended fix, ready for the next shift

1. Pre-orient foam in a line-side bin so the operator grabs and places in one motion. Best-run evidence (7.9 s) shows the step already runs ~5 s faster when foam is positioned well.
2. Move the foam bin inside the reach envelope at the packing station to remove the turn-and-retrieve.
3. Re-capture after the change to confirm step 4 holds near best-run and the variation collapses.

What the fix is worth

Closing step 4 to its best run removes ~5 s per unit. Illustrative, at one line running ~500 units/shift, 2 shifts, 250 days:

TurboProc reports the same three ways every time, throughput, labor efficiency, and cost, drawn from your own session data. When a line is the plant constraint, recovered seconds convert to sellable output, which is why the contribution figure is far larger than the labor figure.