Legal AI Ops Trust Overview

Precision, lineage, and defensibility across legal research and contract workflows.

Demo dataset (guided vertical · mock aggregates)Legalorg_demo_legalUpdated 12:40:48 PMMission Control Stage metrics (Swarm)Task trust replay

← Dashboard

Use this workspace to inspect trust movement across legal reasoning, structured review, and policy-bound outputs.

Contextual focus

Citation lineage
Contract review drift
Policy-sensitive variance

Pipeline KPIs below use the same org aggregates as other workspaces; focus labels describe how to read them for this vertical. Funnel and loss panels use the vertical stage names while the underlying stage keys (generation / reflex / postprocess) are unchanged.

Total tasks

18200

Accepted outcomes

14283

78.5%

Tasks with recovery

5096

Avg lineage / reflex signal

72.0

0–100 scale

Avg time to outcome

4.8 min

Trust Funnel

Reached = corresponding entered*At present on the task aggregate.

Total18200Citation lineage1780097.8%Contract review drift1590087.4%Policy-sensitive variance1494682.1%Outcome1464780.5%Accepted1428378.5%

Citation lineage reach (of total): 97.8%
Contract review drift reach (of total): 87.4%
Policy-sensitive variance (of total): 82.1%
Accepted rate: 78.5%
Contract review drift after Citation lineage: 89.3%
Policy-sensitive variance after Contract review drift: 94.0%

Failed 3553 · Abandoned 364

Recovery Summary

Tasks touched by recovery5096
Recovery rate34.0%
Recovered tasks that converted81.0%
accepted after recovery ÷ tasks with recovery
Avg recoveries per touched task1.40
Most used interventionreflex_retry

Stage Loss Panel

Failed or abandoned tasks counted once at deepest entered stage.

468
Citation lineage
2651
Contract review drift
779
Policy-sensitive variance

Top loss: Contract review drift

Velocity Summary

Averages are per attempt where attempt counts exist on aggregates.

Avg · Citation lineage4.2 sec
Avg · Contract review drift11.8 sec
Avg · Policy-sensitive variance2.1 sec
Avg time to outcome4.8 min
Median time to outcome4.0 min

Batch Impact

Batch impact reflects currently attributed runs only. Click a row for detail.

Run	Targets	Recovered	Accepted after recovery	Avg Δ reflex	Avg postprocess	Failed	Skipped
mock_batch_001	240	186	162	+14.2	1.9 sec	22	8
mock_batch_002	180	142	128	+11.8	2.1 sec	15	5

Mission Health

Sorted by task volume. Open Mission Detail for triage or Swarm for the live board.

Mission	Tasks	Accepted	Failed	Recovery rate	Avg reflex	Top loss	Avg TTO	Actions
mission_demo_primary	4200	3100	680	41.0%	71.0	Contract review drift	4.9 min	View mission View tasks