Skip to content
Request a demo
PlatformModulesIndustriesSecurityPricingInsights Request a demo

The AI engine

Trunnion AI.
The nervous system of the platform.

The native AI engine embedded into every module, dashboard, and workflow in Cortrova - not a bolt-on. A hierarchy of 65 agents reads live operational data across 17 departments and synthesizes it from the shop floor up to the C-suite, predicting problems instead of just reporting them.

Request a demo Explore the platform
65 agents3 tiers17 departments90-day windowSub-1.2s latency
65
Embedded AI agents
17
Departments covered
< 1.2s
Recommendation latency
6 hr
Full analysis cycle

Agent architecture

65 agents. Three tiers. One brief.

Trunnion AI runs a hierarchical agent system that processes operational data through a bottom-up synthesis pipeline - department signals roll up through division aggregators to an executive layer that produces a single unified briefing.

Tier 01 51 agents

Department agents

Three agents in every one of the 17 departments. An Analyst watches live records for anomalies, an Advisor writes per-record recommendations, and a Reporter rolls up KPI summaries - so intelligence lives inside the tools each team already uses.

Tier 02 7 agents

Division aggregators

Seven cross-department agents synthesize the floor up a level - Operations, Safety, Finance, Supply Chain, People, Commercial, and Technology - correlating signals between departments that single modules can never see on their own.

Tier 03 6 agents

Executive agents

Five executive specialists - Strategy, Operations, People, Risk, and Market Intelligence - plus a Synthesizer that fuses every tier into one unified C-suite briefing in roughly 30 seconds.

51 department + 7 division + 6 executive (5 specialists + 1 synthesizer) = 65 agents.

Analysis engines

Every capability, included.

✓Anomaly detection - Z-score detection at ±1.5 sigma surfaces the records drifting out of normal before they become a finding - across every department, continuously.
✓Predictive alerting - Linear-regression trend scoring projects where a metric is heading and raises the alert while there is still time to act, not after the report is filed.
✓Cross-department root cause - Correlation detection across a 90-day rolling window traces a quality escape, an overdue work order, or a schedule slip back to its true source in another department.
✓Prescriptive actions - A library of actionable templates turns each finding into a concrete next step - with an ROI projection and an assigned owner attached.
✓Scenario modeling - Pre-seeded causal relationships let you run what-if scenarios - change a constraint and see the projected ripple across cost, schedule, and risk.
✓SPC monitoring - Continuous Cpk and Cp monitoring runs around the clock, watching process capability so quality drift is caught at the chart, not at the customer.

How it works

A bottom-up synthesis pipeline.

Every cycle reads the floor, correlates across divisions, synthesizes a prioritized briefing, then lands prescriptive actions - and reruns every 6 hours plus on demand.

Read the floor

Department agents read live records across all 17 departments on a rolling 90-day window - no exports, no nightly batch.

Correlate across divisions

Division aggregators connect signals between departments to find the patterns that no single module exposes.

Synthesize the brief

Executive agents and the Synthesizer fuse every tier into one prioritized briefing in about 30 seconds.

Act and verify

Prescriptive actions land with ROI and an owner; a full cycle reruns every 6 hours, plus on demand.

Control & governance

You decide where the engine runs.

Trunnion AI is included with the platform and built to sit inside your data boundary - with instant kill switches, model flexibility, and cloud, on-prem, or air-gapped deployment.

Granular kill switches

Disable AI at global, division, or team scope with instant effect - you decide exactly where the engine runs.

Model flexibility

A model-abstraction layer lets different departments run different models at the same time, matched to the task and your data-boundary rules.

Everything included

All 65 agents, every analysis engine, and every signature capability ship with the platform - no add-on tier, no per-agent cost.

Deploy anywhere

Run the engine in the cloud, on-premises, or fully air-gapped, so AI stays inside your ITAR and CMMC data boundary.

Runs inside your compliance boundary

Built in, not bolted on.

ISO 9001:2015ISO 14001:2015AS9100 Rev DIATF 16949FDA 21 CFR Part 11ITAR (22 CFR 120-130)CMMC 2.0 Level 2OSHAEPAFAR / DFARSDCAA cost accounting

FAQ

Questions, answered.

What is Trunnion AI?

Trunnion AI is the native AI engine embedded into every module, dashboard, and workflow in Cortrova - not a bolt-on analytics product but the nervous system of the platform. It runs 65 agents in three tiers across all 17 production departments. Trunnion AI is also the parent company behind Cortrova.

How are the 65 agents organized?

In three tiers: 51 department agents (three per department across 17 departments - an Analyst, an Advisor, and a Reporter), 7 division aggregators that synthesize across Operations, Safety, Finance, Supply Chain, People, Commercial, and Technology, and 6 executive agents - five specialists plus one Synthesizer that combines every tier into a single C-suite briefing.

What can the engine actually do?

Anomaly detection, predictive alerting, cross-department root-cause analysis, prescriptive actions with ROI projections and assigned owners, scenario modeling against pre-seeded causal relationships, resource-optimization prioritization, and continuous SPC monitoring of process capability - all running on live operational data, not exports.

How fast is it?

Recommendations surface in under 1.2 seconds, an executive synthesis briefing is produced in about 30 seconds, and a full bottom-up analysis cycle runs every 6 hours plus on demand - all against a rolling 90-day data window across 17 departments.

Can we control or turn off the AI?

Yes. Granular kill switches let you disable the engine at global, division, or team scope with instant effect. A model-abstraction layer lets different departments use different models simultaneously, and the whole engine can be deployed in the cloud, on-premises, or fully air-gapped to stay inside your compliance boundary.

Is Trunnion AI an extra cost?

No. All 65 agents, every analysis engine, and every signature capability are fully included in the flat annual rate with unlimited users - nothing about the AI engine is gated or priced separately.

See it live

Watch Trunnion AI read your floor.

Request a demo and see the 65-agent engine surface anomalies, root causes, and prescriptive actions on your own operational data.

Request a demo Explore the platform