Loading…
Technical explainer
Decionis for Google Cloud / Vertex AI: governed execution without losing human approval
Decionis can expose a Vertex AI Extension or ADK tool surface while Decionis Protocol remains the execution-time decision authority. It stays cloud-agnostic, can verify Google identity at the adapter edge, and preserves policy-gated execution, human approval paths, Decision Dossiers, validation packs, and audit evidence before action proceeds.
This is an architectural fit statement, not a claim of official Google endorsement, certification, or native partner status. The point is simpler: teams can use Google Cloud and Vertex AI for intelligence while keeping execution governance intact.
Where Vertex AI fits
Vertex AI can power models, agents, retrieval, and automation logic. That does not remove the need for an execution-time governance layer once a recommendation is about to trigger spend, approvals, routing, or another consequential action.
Where Decionis fits
Decionis sits between intent and execution. It takes the live context and institutional policy, then returns one confidence-scored governed verdict: execute, block, escalate, or deliberately hold when evidence or confidence is insufficient.
Why cloud-agnostic matters
Decionis governs the decision boundary itself, so it can sit in front of Google Cloud and Vertex AI workflows without forcing the rest of the operating stack to move. Teams, Power Automate, ERP, Salesforce, or ServiceNow can remain the workflow surfaces.
What is already implemented
The Vertex adapter can already verify Google identity at the edge, forward intent into Decionis Protocol, and return a governed verdict with Decision Dossier proof plus verified_decision metering metadata.
Execution control retained
What Decionis preserves around a Google stack
- Policy-gated execution before downstream action
- Human approval paths when policy or confidence requires it
- Decision Dossier and verification envelope for each governed action
- Validation packs with integrity hashes for audit and supervisory review
- Org-wide and per-decision kill switches for controlled rollback
Why this matters for enterprise trust
Different from broad IRM and GRC platforms
ServiceNow IRM and IBM OpenPages are broader governance platforms. Decionis is narrower and closer to the point of action: it returns the governed verdict before execution proceeds.
Different from LLM guardrail tooling
Guardrails AI, Lakera, and LangChain Guardrails constrain models, prompts, outputs, or AI application behavior. Decionis governs whether a business action should execute now under institutional policy.
Different from a services-led integration
The same public docs and self-serve workspace can onboard a partner, issue credentials, manage versions, update routes, and export validation packs without turning the rollout into consulting disguised as product.
Enterprise hardening path
Google OIDC / WIF
Adapter-edge Google identity verification is available now and can move from optional to required per deployment.
Observability
Structured logs emit project, location, subject, verdict, and metering labels so Cloud Logging and Monitoring can ingest them cleanly.
Metering
The adapter already returns a verified_decision metering unit so downstream billing and reporting can attach to a stable control-plane event.
PSC / regional / CMEK
These remain enterprise deployment hardening tracks. They strengthen the Google posture without changing the Decionis Protocol decision boundary.
FAQ
How does Decionis fit into a Google Cloud or Vertex AI stack without losing governance or human approval?
Decionis fits as a thin Vertex AI Extension or ADK tool surface that forwards agent intent into Decionis Protocol for the actual governed decision. The adapter can verify Google identity at the edge, then return one governed verdict, Decision Dossier proof, and billable verified_decision metadata while Google Cloud and Vertex AI stay in the distribution and intelligence layer.
How does Decionis stay cloud-agnostic while Microsoft Teams and Power Automate remain workflow surfaces today?
Teams and Power Automate are workflow surfaces, not the permanent home for policy. Decionis governs the execution decision itself, so it can sit in front of Microsoft, Google Cloud, Vertex AI, ERP, and other stacks while keeping the same proof and approval model.
Can Decionis work with Google Cloud or Vertex AI while preserving the current Microsoft Teams, Power Automate, and Decision Dossier proof story?
Yes. Decionis can govern Google Cloud or Vertex AI driven execution while Teams and Power Automate remain the current collaboration and workflow surfaces. The Decision Dossier, verification envelope, validation pack, human-approval posture, and enterprise hardening path for PSC or regional/CMEK controls remain aligned.
The operating model stays stable
Teams and Power Automate can remain the current workflow surfaces. Google Cloud and Vertex AI can sit behind the governed execution boundary. Decionis keeps the same proof story intact: one confidence-scored verdict, one Decision Dossier, one validation-pack layer, and explicit human-approval and rollback controls when they are required.