A
AiNOS x Colosseum
Ontology Decision Twin
Recommended call
Monitor — no intervention required
Separated source systems are connected enough for replay, but no incident pressure exists yet.
Why you can trust this callSource → ontology → bounded decision
CTX-T00What the rules found0 overloaded resources · 4 stale sourcesdeterministic selectors · 0 trace gaps kept · RULE-BOOTSTRAP-TOPOLOGY
Evidence pulled in27 nodes · 48 edgesbounded graph expansion · 12 edge families
Why this call, not anotherPack line 01 53%top constraint driving the recommendation
How the model was bounded
Rerank candidate nodes by whether they explain the active decision, not by visual adjacency alone.
Conflicts preserved, not hidden
labor_shift_roster.csv sync is watch with 13m lagamr_state_snapshots.jsonl sync is watch with 10m lagtms_outbound_shipments.csv sync is watch with 13m lagexception_and_action_events.jsonl sync is watch with 10m lag
Ask the model to explain this call
A bounded LLM reads only this decision packet — no free reasoning — and returns why it was chosen, what it rejected, and what a human must approve.
Source evidence chainCHAIN-T00
Source evidence3 nodes
Derived riskDEC-T00
Context packet640 tokens
Separated source systems are connected enough for replay, but no incident pressure exists yet.
vault_deployment_context.jsondeployment_id=VAULT-COL-SEL-FC-01integration cadence: WMS daily export; WCS event export; AMR partial realtime; TMS manual exportconfiguration
facility_locations.csvlocation_id=CTRL-ROOM-01 · source_location_code=OPS-CTRLfacility topology: 62 locations with restricted cold-zone aliasesstatic master
automation_assets.csvasset_id=AMR-11fleet inventory: 8 AMRs, 2 chargers, 1 pack stationstatic master
Focus entities
CTRL-ROOM-01AMR-11AMR-17AMR-71SEL-FC-01
203 source rows omitted after event-source filtering and entity focus compressionHow the ontology expanded to this call27 nodes · 48 edges traversed
Open in atlas IncidentRiskDecisionSourceRecord
All 18 bounded nodes
SourceRecordsource_system_sync_runs.csv · SYNC-T00-01source-grounded
SourceRecordsource_system_sync_runs.csv · SYNC-T00-02source-grounded
SourceRecordsource_system_sync_runs.csv · SYNC-T00-03source-grounded
SourceRecordlabor_shift_roster.csv · LAB-T00-ROLE-C...COLD-02
SourceRecordlabor_shift_roster.csv · LAB-T00-ROLE-DOCKDOCK-03
LocationCTRL-ROOM-01 · control roomCTRL-ROOM-01
WorkTaskTASK-00002 · pick_movePACK-01
WorkTaskTASK-00006 · pick_movePACK-03
WorkTaskTASK-00007 · pick_movePACK-02
WorkTaskTASK-00011 · pick_movePACK-04
WorkTaskTASK-00012 · pick_movePACK-04
WorkTaskTASK-00014 · pick_movePACK-03
WorkTaskTASK-00022 · pick_movePACK-01
WorkTaskTASK-00028 · pick_movePACK-01
AssetAMR-11 · LIFT-250RSV-01
IncidentT00 · Facility baseline establishedCTRL-ROOM-01
RiskT00 operational pressureCTRL-ROOM-01
DecisionDEC-T00 · observeCTRL-ROOM-01
Evidence & simulation detailProbe the call — what-if alternatives, telemetry, and capacity stress.
Simulation telemetryMon 09:05
Throughput68units/hr
Task queue12open
AMR util.42%fleet
Source lag3mp95
Inventory trust98%confidence
Dock slack180mremaining
Pressure
Queue
Source lag
Resource capacity stressPack line 01
Peak util.53%
Overloaded0
Backlog0
Simulation causal replayBOOTSTRAP -> T00
Pressure8+3
Queue12+2
Dock slack180m0m
Lag3m+1m
RULE-BOOTSTRAP-TOPOLOGYLoad facility topology, AMR master data, and integration cadence before incident replay.Baseline state has source separation but low operational pressure.
baseline_order_mix22 orders / 38 linesfixed account/service-tier mix with bounded SKU variety
scan_upload_jitter3m p95 laguniform 0-5 minute ingest delay
No operational incident is active; the replay starts from a plausible separated-source warehouse state.