Custom Software
Your workflow, built on the Trunnion control plane.
Custom builds bring the same governed approval, audit, and deployment patterns that power the productized portfolio to bespoke client systems: purpose-built governed AI workflows, SaaS products, internal platforms, integration layers, and automation systems.
Custom Software engagements deliver purpose-built governed AI systems on the Trunnion control plane: your data, your policy, your models, your deployment target.
It is designed for policy-based authorization, layered tool controls, reviewable execution records, and human-in-the-loop approval gates, with implementation and tests verified for the exact release, deployed to cloud, on-premise, or fully air-gapped environments (security boundaries). Trunnion AI is aDuskbridge company, built by the team behind Viceroy NM.
Feature highlights.
The capabilities that carry the custom & labs workflow in Custom Software, on the governed Trunnion control plane: agents draft and route the work while a human operator stays in command.
Representative feature highlights. Actual screens vary by deployment and configuration.
Organizations that need a purpose-built governed AI workflow, SaaS product, internal platform, integration layer, or automation system.
What Custom Software does.
Built on the proven Trunnion control plane, not from scratch
Your data, your policy, your models, your deployment target
ABAC, human-review gates, and reviewable audit records scoped and verified per release
Cloud, VPC, on-premise, or fully air-gapped delivery
Fixed-scope or milestone-based engagements
Inside Custom Software, capability by capability.
Built to your business
Identity, approvals, audit, model behavior, and routing are scoped, implemented, and verified for the contracted release rather than assumed from a shared pattern.
Your domain, modeled
The work graph is built around your records, whether work orders, routes, bins, or solicitations, with typed edges and full field provenance.
A proven pattern
A custom vertical can reuse reviewed platform patterns, subject to release-specific architecture and control validation.
Composition, not a blank page
A new vertical is the invariant core, a domain object model, a domain skill pack, and seeded knowledge bundles.
Deployment is configuration
The same stack ships cloud, VPC, on-premise, and air gap; changing the boundary is a configuration change, not a rewrite.
Governance scoped for the release
Each custom build defines and verifies its authorization, approval, audit, and deployment controls in the statement of work and release evidence.
Attribute-based access control
Attribute and role-based authorization can be configured for a deployment; the policy model and tests are supplied as release-specific evidence.
Human-in-the-loop gates
A permissioned human accepts, edits, or rejects consequential outputs, on approval paths computed from role, attribute, and policy.
Reviewable audit record
The release is designed to record policy decisions, model actions, tool calls, and approvers. Cryptographic linkage, coverage, and replay are verified for that release.
Guardrails in code
External sends, money movement, permissions, and grounding controls are defined and tested for the specific workflow; no universal prevention claim is made here.
Efficient and self-improving
The router sends every request down the cheapest capable path, and every human decision makes the system smarter.
Model routing as policy
Deterministic rules run first, ML models second, language models last, and model choice is configuration, never code.
Cost governance built in
Token usage is recorded per call, departments get monthly AI budgets and per-action caps, and expensive jobs batch off peak.
Decisions become training data
Every accept, edit, and reject is captured as label data that retrains models and improves skills over time.
Knowledge that travels
Curated knowledge bundles seed a new deployment on day one, carrying permissions, provenance, and classification with the data.
What Custom Software is for, and what it is not.
Scoped per engagement: the intended workflow, decision boundaries, data expectations, and human review points are defined in the statement of work before the build starts, and documented with the delivered system.
Like every Trunnion product, Custom Software is designed to propose, draft, and prepare work for human decision, with consequential actions pausing at a human approval gate. AI outputs may be inaccurate, incomplete, or unsuitable and should not be the sole basis for a consequential decision affecting a person. Expected inputs, model providers, known failure modes, and the exact human review points for a deployment are documented in the release documentation and written agreement. See theAI transparency notice and theAI acceptable use policy.
Custom Software: key facts.
| Product | Custom Software |
|---|---|
| Category | Custom & Labs |
| Status | Custom build |
| Platform | Trunnion control plane: governed multi-agent orchestration |
| Governance | Designed for policy-based authorization and layered tool controls; verified per release |
| Audit | Reviewable execution records; integrity and replay verified per release |
| Human oversight | Human-in-the-loop approval gates on consequential actions, by design |
| Model posture | LLM-agnostic model routing, hosted or on-premise |
| Deployment | Cloud, on-premise, or fully air-gapped |
| Classification | Designed for classification-aware workflows up to TS/SCI; accreditation determined per environment |
| Security alignment | Engineered to align with NIST SP 800-53 control families; alignment by design, ahead of formal certification. |
| Developer | Trunnion AI, LLC, a Duskbridge company |
Governed by design, like every Trunnion product.
Custom Software inherits the platform's governance, audit, and deployment patterns instead of rebuilding them: policy on every action, a human operator in command, and a record you can hand to an auditor.
Policy-aware governance
Attribute-based access control (ABAC) is designed to decide what every agent may see and do, per tenant, role, and classification.
Human-in-the-loop gates
Consequential actions are designed to pause for a named operator to approve, edit, or reject before anything executes.
Reviewable execution records
The architecture is designed to record actions, approvals, and supporting traces for review. Cryptographic implementation is verified per release.
Common questions about Custom Software.
Are we starting from zero on a custom build?
No. A new system composes the proven control plane with your domain object model, your skill packs, and seeded knowledge bundles, built to your needs, so governance, audit, ML, and routing arrive already built.
Are we locked into one AI provider?
No. Mission logic stays separate from model choice. The router runs across approved hosted models and local runtimes, and nothing outside the router references a provider by name.
Can the AI act without a human?
Consequential actions hold at human gates, and hard guardrails are enforced in code. Autonomy is progressive, moving from supervised pilot to gated independence as evidence accumulates.
Next step
See Custom Software against your workflow.
Request a demo and see the governed control plane run in your environment, from cloud to air-gap.