# The prompt sent to the agent (amr)

```text
I'm the proposal engineer at Brindlefield Automation. Last week our AI built Corrin & Vale's Unit 4 warehouse in Laydyne and we proposed Option B: the four packing benches moved south of the pick zones, and three robots instead of four. That project is open in the Laydyne Studio. The live layout is the customer's current layout; scenario cv-option-a-4 is the peak on it (110 orders an hour, 20 runs from seed 20261006); and our two hand-made options are saved as layout options "Option A — current layout — 4 BF-T1" and "Option B — south packing benches — 3 BF-T1".

Before we send the final proposal I want Laydyne to search the options instead of us picking two by hand:
- What may change: the number of robots (the scenario's parameter robots: 2, 3, 4 or 5); the four packing benches pack-1 to pack-4 as one block, with its centre anywhere on the aisle line z = 27.5 m from x = 5 to x = 53 m in 2 m steps; and the spacing of the benches (4, 5.5 or 7 m centre to centre).
- Keep clear: the office (x 0.38, z 31.92, 11.55 × 7.7 m) and the outbound staging lane (x 39.5, z 33, 19 × 3.8 m). No new layout-check problems.
- The peak must still reach 110 orders an hour.
- Goal: the fewest robots; among those, the lowest average wait.
- The same 20 runs and seed as the proposal, so the numbers line up with Option B.

Use search_layouts. Then compare the best with our Option B and with Option A using compare_process_options on the saved options, and tell me which differences are beyond the noise. Show me the best layout: save the current layout as design option "Customer layout", apply the best candidate, save it as design option "Search best", show me a 2D plan with the differences from the customer layout and a 3D view, then restore "Customer layout".

Finish with a short table for Option A, Option B and the search's best: robots, where the benches are, throughput, average wait and robot distance, with 95 % intervals. Say it is the best of the candidates searched, not a proven optimum, and what the model does not calculate.
```