AlterspectiveSharedo Audit

Tier 3 · Lifecycle

Work type health & anomalies

Configured work types are a hypothesis about the work; the matters are the evidence. When one type absorbs a third of the book, it is usually hiding several distinct kinds of work that deserve their own routing, reporting, and resourcing. This audit reads the titles and the lifecycle to find the mismatch.

T3Intelligence

1What we examine

roadmapConfigured taxonomy versus realityCFG-WT

On-premise embedding analysis of matter titles segments a bloated configured type into the linguistically distinct kinds of work living inside it.

Phase-plan defectsCFG-PH-01 · CFG-PH-12

Multiple start phases, unreachable phases, and a missing mandatory Draft phase — structural defects that confuse routing and reporting.

Inheritance depthCFG-WT-01

Work-type inheritance beyond the sane limit (the five-level limit with the multi-dimension test) — depth that makes change risky.

Guard anomaliesCFG-PH

Phase guards that block legitimate transitions, or are absent where a control is needed.

Stuck queues & velocityCFG-PH

Reopen and pending-close phases holding years-old matters, and close velocity across the book.

2How we examine it

Matter titles are embedded with a local model on an on-premise stack — client data never leaves a controlled environment — and segmented; configuration lenses check the phase plans, guards, and inheritance depth. Segmentation is exploratory and reviewed with your team.

Cluster segmentation is exploratory, not a discovered 'true' taxonomy: the number of segments is a judgement, and the labels are model-derived inferences reviewed with your team — as the merge above shows, that review sometimes finds two proposed segments are really one. Counts and phase ages are observed.

3Example finding

Illustrative example — synthetic demonstration corpus

One configured type absorbed 41% of the book (12,231 of 30,000 matters). Title embeddings, computed on-premise and labelled by a local model, segmented it into 7 distinct kinds of work (two of the eight AI-proposed segments read as the same kind of work and were merged before review — the audit surfaces a candidate list, a human confirms it).

  • 7 linguistically distinct kinds of work hidden inside a single configured type
  • Two phases (Re-Open, Pending Close) holding matters 2–8 years old
  • Phase-plan and inheritance defects flagged against CFG-PH and CFG-WT rules
41%one type
Hidden inside one typeRest of the book
First Party Property Damag3,248 mattersDirectors & Officers Liabi1,653 mattersCyber Incident Response1,605 mattersWorkplace Injury Claim1,503 mattersPublic Liability Personal 1,479 mattersMotor Vehicle Personal Inj1,387 mattersMedical Negligence Defence1,356 matters
See it live — this is a static snapshot of one finding. The underlying Matter Universe renders all 30,000 synthetic matters as one interactive map — drag to orbit, click any region to have it name itself, and see exactly why a given matter sits where it does. In the sample report it has a full chapter — how to read the map honestly — and the taxonomy breakdown behind this finding is browsable in full.
Left: the share of the book sitting under one configured type. Right: the 7 distinct kinds of work the title embeddings found inside it.

4The benefit

What you walk away with

A taxonomy that matches the work you actually do — sharper reporting, routing, and resourcing — and a phase-plan defect list that unblocks the matters sitting in stuck queues.

← All audit domains See it in a full report