ONTOGUARD
See OntoGuard in Action Talk to Us
Decision Authorization Headless Runtime Financial Services Insurance Healthcare Payers Cybersecurity Vertical AI Vendors Implementation Partners Proof Explorer Financial Proof Case Technical Diligence Integrations & Partners Resources About Privacy Contact
  1. OntoGuard
  2. Resources

Resources

Videos, proof assets, and concepts.

Product video

See Decision Authorization in action

1:05 · Decision Authorization demo

Controlled, buyer-safe demonstration of proposed action → decision → proof. Does not autoplay.

Public proof assets

Anonymous public evidence is limited to buyer-facing artifacts. Deeper seam and audit packs are available through technical diligence.

ZIPPublic Decision Authorization Evidence PackDownload
PDFDecision Authorization overviewDownload
PDFFinancial services one-pagerDownload
PDFFinancial Decision Authorization packetDownload
JSONPublic decision receiptOpen
JSONPublic evidence manifestOpen
HTMLInteractive Proof ExplorerOpen
MAILTechnical diligence proof packRequest proof pack

Research

September 2026 preprint

Decision Authorization for Trustworthy AI: Semantic Institutional Admissibility in a Neurosymbolic Control Plane — Reproducibility and Disclosure Scope

Founder-authored preprint defining Decision Authorization as a distinct control problem. Preprint; not represented as peer-reviewed publication or independent product validation.

Read the preprint ResearchGate profile

Decision Authorization concepts

Decision Authorization resources

ALLOW / BLOCK / ESCALATE is the product language. A score is evidence, not authorization.

ESCALATE routes the action to qualified human authority and records reviewer, evidence considered, disposition, release state, and resulting consequence.

No-bind proof records that a protected effect did not form.

Ontology supplies structured semantic context so the system can interpret the attempted change. It is not the commercial category.

FAQ

Is OntoGuard a monitoring or guardrail product?

No. Monitoring observes. Guardrails constrain. OntoGuard determines whether an exact proposed transition may proceed and proves what happened afterward.

Do I need to upload internal documents to start?

No. Start with representative actions. Do not send confidential customer data through the public form.

Does the site claim production L5 everywhere?

No. Controlled proof is demonstrated. Customer-route production completeness requires integration evidence.

Is the AgenTrust integration Verified?

It is an accepted Community Integration. Verified is a separate tier and is not claimed.

Does OntoGuard replace human authority?

No. ESCALATE exists specifically to route to qualified human authority.

Transcripts

Decision Authorization demo (1:05)

AI agents can now move money and change records. The question is whether this exact action is authorized to happen. OntoGuard sits at that decision boundary. It evaluates the exact proposed action against evidence, governing basis, intent, and authority. The result is explicit: ALLOW, BLOCK, or ESCALATE to qualified human authority. Change something material — and the prior decision no longer binds. Authorization binds to the action that was actually evaluated. OntoGuard records the decision: release state, human-review requirement, and whether the protected consequence formed. A self-contained headless layer inside the enterprise execution environment. AgenTrust Marketplace — Community Integration. Twenty-seven of twenty-seven adapter tests passed in maintainer reproduction. OntoGuard. Decision Authorization for Agentic AI. ALLOW. BLOCK. ESCALATE.