Ghost Citadel
Menu
PlaygroundPlay

Ghost Observatory / Industries

Your operation is bigger than its org chart.

Coming soon

Follow production orders, shipments, claims and service requests through their recorded steps. Find delays, repeated work and control breaches, then inspect the records behind them.

Explore six industries See the software

FROM OPERATIONAL DATA TO DECISIONS.

  1. 01 Read the records from your systems
  2. 02 See the process and its delays
  3. 03 Trace the findings and prioritise action

Every finding connects to its source records.

A worked claims example.

Inspect a complete result from synthetic claims data. Trace the rule, the records it could check and the claims it flagged.

Explore the published sample
01 / 06 — Manufacturing

Find the capacity your process wastes.

Approvals, repeated inspections and late repairs consume capacity. Follow production orders through the full sequence to see where time is lost and where improvement can start.

See the case for bringing work home

Example sequence

  1. Production order
  2. Build
  3. Quality check
  4. Rework / release
Question, evidence and decisions
Ask
Which production orders return to quality checks, and how much of the recorded lead time sits between operations?
Inspect
Event sequences, repeated activities, elapsed time between recorded steps, and team assignments.
Decide
Use the findings to improve inspection capacity, production sequencing or work allocation and strengthen the case for competitive local production.
Needs
Production-order ID and activity; declared event order for paths and timestamps for timing analysis; duration and team if available.

Explore manufacturing process mining: data, rework and the reshoring decision →

02 / 06 — Logistics & supply chains

Keep orders moving.

Follow orders across fulfilment systems to find the handovers and repeated handling that stretch delivery times. Focus improvement on the steps that hold up the customer’s order.

View process and timing analysis

Example sequence

  1. Order released
  2. Pick
  3. Pack
  4. Dispatch
Question, evidence and decisions
Ask
Which handoffs account for the longest delays, and which orders cycle back through the same handling step?
Inspect
Recorded order paths, transition timings, repeated handling steps, and unresolved records.
Decide
Target changes to staffing, routing, escalation rules or the way teams hand over work.
Needs
Order or shipment ID, steps, ordering information, and recorded times; assignees if available.
03 / 06 — Banking & insurance

Make every approval accountable.

Trace claims and payments through assessments, approvals and exceptions. Check who performed each responsibility and identify the records that need a closer review.

Inspect the synthetic claims example

Explore a separation-of-duties example

Example sequence

  1. Claim received
  2. Assessment
  3. Approval
  4. Settlement
Question, evidence and decisions
Ask
Did the same person assess and approve a claim? Which claims cannot be checked because the record is incomplete?
Inspect
Actor identity per step, ordered activities, explicit separation-of-duties rules, and records missing a required person or step.
Decide
Investigate every breach of the declared control and close the gaps in recordkeeping.
Needs
Claim ID, activity, assessor/approver identities and the declared rule; declared event order for paths and timestamps for timing analysis.

Published example: synthetic claims data. Inspect the 300-claim result →

04 / 06 — Healthcare administration

Give staff more time for essential work.

Find the delays and repeated handovers in referrals, appointment scheduling and administrative follow-up. Use the workload evidence to improve routing and make better use of staff capacity.

See the Reader views

Example sequence

  1. Referral
  2. Triage
  3. Scheduling
  4. Follow-up
Question, evidence and decisions
Ask
How long do referrals remain between administrative steps, and how often are they reassigned?
Inspect
Documented referral journeys, transfer counts, delays, and repeat administrative work.
Decide
Improve referral routing and scheduling capacity based on the administrative workload the records show.
Needs
Referral/work-item ID and activity; declared event order for paths and timestamps for timing analysis; administrative team or staff identifiers where appropriate.
05 / 06 — Government & public services

Shorter backlogs. Stronger accountability.

Follow applications from submission to decision. Identify where work accumulates, which service targets are missed and which repeated steps need attention.

See investigations and required data

Example sequence

  1. Submission
  2. Review
  3. Decision
  4. Notification
Question, evidence and decisions
Ask
Where do applications accumulate, which ones miss declared service targets, and which steps repeatedly send work back?
Inspect
Submission-to-decision paths, service-time measures, reopened work and documented rule exceptions.
Decide
Choose where to add processing capacity, redesign a handoff or investigate inconsistent handling.
Needs
Application ID and process activities; declared event order for paths and timestamps for timing analysis; target rules and status where relevant.
06 / 06 — Energy & utilities

Reduce repeat visits. Improve completion.

Track work orders from fault report to verified closure. Find recurring visits, reopened jobs and approval delays so you can improve maintenance and field service.

See supported data requirements

Example sequence

  1. Fault reported
  2. Work assigned
  3. Site visit
  4. Verified close
Question, evidence and decisions
Ask
Which maintenance work orders reopen, and where do they wait between assignment, intervention and closure?
Inspect
Work-order transitions, reopen events, waiting intervals and team handovers.
Decide
Use the findings to improve dispatch, parts readiness, service sequencing and closure controls.
Needs
Work-order ID, activities and close or reopen state; declared event order for paths and timestamps for timing analysis; crews and duration if recorded.

Run it locally

Serious analysis. Everyday hardware.

The complete Ghost Observatory Engine and Reader run locally on a MacBook Neo or a Raspberry Pi 5. Analyse your exports, explore the SQLite result and keep your data and infrastructure under your control.

No operational data is sent to Ghost Citadel.

The hardware story See installation
EVERYDAY LAPTOPMacBook
Neo.Local Engine + Reader
SMALL ARM COMPUTERRaspberry
Pi 5.Local Engine + Reader

One result file. No mandatory cloud account.

Start with your actual process

Tell us which process needs to improve.

Tell us where work slows down, which systems record it and what you want to improve. We will identify the exports and analysis that support your next decision.

Discuss your operation An End to Outsourcing

Choose the evidence for your industry.

Industry labels do not change the meaning of the records. A production order, shipment or service request needs a defined identity, recorded activities and evidence for the measure you want.