Claude reviews the already-computed tier results and explains each issue in plain English, telling you exactly what to fix in the source sheet. No figure is recalculated — it works from the audit output only.

The data is mostly complete and usable, covering 6 of 7 files and 101 of 106 machines for one confirmed month, but several gaps and flags need attention before the figures can be fully trusted. Key missing items include no data at all from the HDPE daily workbook for June 2026, five PTMT machines with no recorded runs, and planned-hours baselines not set for 12 machines across Moulding and Pipe, which means utilisation and efficiency figures for those machines are unreliable. You should also verify the logged run hours for two machines showing utilisation above 100 percent, confirm whether the low output figures for Moulding A01, PTMT Grinder-2, and Garden M/C-1 are genuine or recording errors, and note that Tank and Tank variant outputs are shown without utilisation or efficiency since those lines record items only with no machine or run-hour data.