Case Library
Action Register
Every corrective and preventive action across all cases — the follow-up view. Overdue items come first. Click a row to open its case at the Actions step.
Failure-Mode Library
Seeded with common CCPP failure modes. Fully editable — entries are offered as dropdown choices during case creation and as domain hints to the assistant.
Asset Register
Asset ID / Asset combinations, selectable on every RCA case. Import expects two columns on the first worksheet — Asset ID and Asset (header spelling is flexible: e.g. Tag, KKS, ID / Name, Description). Rows are matched on Asset ID: new IDs are added, existing ones get their name updated.
Knowledge Services
Connect a licensed MCP knowledge server that provides predefined Asset Type · Failure Mode · Failure Effect · Recommended Task. A subscription license key is required (billed monthly or yearly). Records can be browsed here and pulled into a case during investigation.
Browse
Administration
Users (whitelist)
Only users listed here can sign in. Add engineers and admins for your plant.
Lists of values
Dropdown options offered across the app — e.g. the Systems list used in filters, the case form and the failure-mode library. Deleting a value never changes existing cases; it just stops being offered. Case status and consequence are fixed because the workflow logic depends on them.
Case templates
Preconfigured entry points offered in the new-case dialog. Five CCPP templates are built in; add your own or override a built-in by using the same key. Org templates seed the case exactly like built-ins: frame fields, a timeline skeleton, and failure-mode-library entries as candidate causes.
Report branding
Align the generated RCA report (on-screen, PDF and Word) to your company standard: a logo at the top and a footer line printed on every page, plus an optional company name. Leave everything blank to use the default look.
Backup & Restore
Download a portable backup of all data (cases, causes, actions, assets, failure-mode library, users, connections). Restore replaces all current data.
Microphone & dictation
Check that the browser can hear your microphone before using the dictation
(mic) buttons in the Effects step. Pick your input device, click Test microphone
and speak — the level bar should move and turn green. The mic works on
http://localhost / HTTPS only, and needs permission for this site and, on
macOS, in System Settings ▸ Privacy & Security ▸ Microphone.
- 1Operating context
- 2Case
- 3Effects
- 4Questions
- 5Investigate
- 6Actions
- 7Report
Quick RCA — the Questions step is skipped and the deep-analysis tools are hidden. Raise the rigor level in step 2 · Case if the investigation grows.
Operating context
Describe the operating context that frames this event — duty, load, run hours, cycling, ambient, fuel, campaign stage — the conditions the equipment was actually working under. This sets the scene for the whole analysis. Save a context you reuse and load it into later cases.
Case details
Consequences & investigation rigor
These facts determine how deep the investigation should go. The recommendation below is rule-based and recorded with the case.
Guided effect capture
Describe what was observed — not what you think caused it. A few keywords per box are enough; the assistant asks follow-up questions next. Tip: dictate or type everything into one box, then let AI sort it into the other fields.
The operating context that framed this event is captured on the first tab.
Event timeline (what happened, in order — approximate times are fine)
Diagnostic follow-up questions
The assistant asks up to 5 targeted questions, like a senior reliability engineer would. Answer what you can — every question is skippable.
Root causes & contributing factors
Potential root causes (drag or use ▲▼ to rank)
Contributing factors
The causal tree is built from the stored causes — link a cause to the one it explains ("Why?") to grow branches. The 5-Whys chains below are read from the same tree, never invented for the report.
5-Whys chains (root-to-deepest path per branch)
Brainstorming view — group hypotheses by category to spot blind areas. Every entry is a normal hypothesis in the shared model and still needs evidence before it can be confirmed; moving one between columns changes only its category.
Barrier analysis (comprehensive rigor — which safeguards failed, were missing, or were bypassed)
Change analysis (what changed vs. what was expected — and was it authorised and assessed?)
RCA diagram (green = confirmed · amber = undetermined · dashed grey = excluded)
Corrective & preventive actions
Draft actions for the confirmed items, then edit and assign owner and due date.