AlterspectiveSharedo Audit

Tier 4 · ODS & Parties

Party & data integrity

The Operational Data Store (ODS) is where the same organisation ends up recorded four times, where connections point at parties that no longer exist, and where a party can quietly sit on both sides of a matter. This audit resolves the duplicates — conservatively — and surfaces the screening signals your conflicts process needs.

T4Intelligence

1What we examine

roadmapDuplicate organisations

Case, punctuation, and suffix variants of the same organisation are grouped — with a deliberate no-auto-merge guardrail that rejects look-alikes rather than risk a wrong merge.

roadmapConnection integrity

Orphaned edges — connections pointing at parties that no longer exist.

roadmapExternal-reference reconciliation

ODS references reconciled against the practice-management system (PMS) to find records that disagree across systems.

roadmapInactive-party hygiene

Parties marked inactive but still connected to live matters.

roadmapActed-both-sides screening

Parties appearing on both the acting and the opposing side — surfaced as screening candidates for your conflicts process, never as conflict determinations.

2How we examine it

Entity-resolution and graph lenses run over ODS parties and connections on an on-premise stack (client and third-party data — full authority gate, controlled environment only). The matching guardrail is deliberately conservative: when two records only look alike, it rejects the merge.

Entity matches are model-derived inferences with a conservative guardrail — look-alikes are rejected, not merged. Both-sides hits are screening candidates for your authoritative conflicts process, never determinations. Duplicate and orphan counts are observed.

3Example finding

Illustrative example — synthetic demonstration corpus — underlying lens on our roadmap, not yet in the productised pipeline

The register held 333 duplicate organisation groups. The guardrail correctly refused to merge 380 look-alike name groups that shared a spelling but carried different company reference numbers — distinct real entities — and the graph surfaced 1,305 acted-both-sides screening candidates.

  • 333 duplicate organisation groups grouped for review
  • 380 same-spelling, different-reference name groups correctly rejected — never auto-merged
  • 1,305 acted-both-sides screening candidates for the conflicts process
Duplicate org groups (merg333Look-alike name groups cor380
333duplicate organisation groups found
1,305acted-both-sides screening candidates
Duplicate organisation groups proposed for merge, and the look-alike name groups the guardrail deliberately rejected (in green — the guardrail working).

4The benefit

What you walk away with

A cleaner party register with no dangerous auto-merges, reconciled external references, and a both-sides screening list your conflicts team can clear — surfaced for a human to decide, not decided for them.

← All audit domains See it in a full report