Every part of the investigation workflow, and the reasoning behind how each one is built.
An investigation names its asset from the same tree every module shares, so the failure history and the machine record are one click apart.
A quick chain for the conveyor jam, a branching tree for the one that stopped the line. The program reads both the same way.
Drag causes into place, pan a big tree from the minimap, and keep the toolbar where you are working.
The technician at the teardown photographs what they found while the part is still in their hand, and the photo lands on the investigation.
Every corrective action names who owns it and when it is due, so nothing dissolves into "we should".
The investigation stays open until the fixes are confirmed in service, not until the report reads well.
A cause carries its support - the photo, the lab result, the document - so the conclusion holds up in front of people who were not there.
A review step before closure, so the conclusion is agreed rather than assumed.
The printable report holds the reasoning, the evidence, and the actions - not a summary of them.
On paper the diagram becomes an indented outline, because a scrolling canvas dies in print and an outline does not.
Downtime and cost rolled up by machine, so the next capital request argues itself.
The same failure on the same machine is a pattern, and the analytics say so without being asked.
MTBF trended, which is the number that proves whether the reliability work is working.
The demo plant has this module switched on, with records already in it.