Open Framework · CC BY 4.0 · Est. Sweden 2025
The Verkflöde Agent Operating Model. An open framework for structuring AI decision authority in enterprise workflows.
The Problem
Enterprise AI has moved from experimentation to operational delegation.
AI agents now influence financial approvals, contract decisions, HR processes, compliance monitoring, and customer operations. Most governance frameworks define what AI is allowed to do in principle. Few define how AI systems exercise decision authority within live workflows.
That gap is where incidents happen. VAOM closes it.
The Framework
VAOM sits between governance intent and operational execution. It translates policy into structured delegation boundaries, confidence thresholds, escalation logic, and audit traceability.
Seven-layer architecture. Six delegation patterns. Composite confidence scoring, calibrated per decision type. Delegation bands with explicit no-automation zones and derived authority limits. Funded and measured human oversight. Delegation identity, credentialing, and leases for continuous agents. Boundary verification and continuous assurance with circuit breakers. Governance of learning and model change. Regulatory design obligations, with a quarterly Alignment Annex mapping the EU AI Act, GDPR, DORA, NIS2, NIST, ISO 42001, and more.
Not a product. Not a platform. A framework any organisation can adopt, adapt, and build on.
Interactive
An interactive browser experience that teaches VAOM through play. Three rounds. Four scenarios. Calibration anti-patterns surfaced in real time as you delegate.
Try Delegation Lab at verkflode.eu →Principles
The principles are not aspirational. They are the load-bearing constraints that make VAOM a framework and not a checklist.
Delegation Authority Matrix
Every AI decision explicit, auditable, and owned by a named human.
| Scenario | High confidence | Medium confidence | Low confidence |
|---|
Delegation Patterns
Agent assembles context and retrieves data for a human decision; it makes no decision and takes no action.
Agent produces a draft decision or recommendation; a human reviews, modifies if necessary, and approves before execution.
Agent classifies incoming items and routes them to the appropriate handler on defined rules and confidence thresholds; it does not resolve them.
Agent makes the decision and executes autonomously within defined authority; humans review a sample of completed decisions after the fact.
Agent operates continuously within defined parameters; humans are alerted when thresholds are crossed and can intervene at any point.
Multiple agents collaborate, each within its own delegation boundaries; escalation to a human occurs when any agent meets a decision outside its authority or chain confidence falls below threshold.
Non-delegable decisions sit outside the six patterns. They route to a named human role; the agent may prepare the brief but never executes the outcome.
Whitepaper
Full specification. Eight principles, seven layers, six worked examples from invoice approval to delegated purchasing. Calibrated confidence gates, derived authority limits, funded and measured oversight, delegation leases, boundary verification, and the delegation scorecard. Ninety-day implementation roadmap, with a companion Alignment Annex revised quarterly.
Who maintains VAOM
Verkflöde AB is a Swedish company that maintains the VAOM framework as an open specification under CC BY 4.0.
The framework draws on two decades of practice in enterprise technology, learning systems, and AI governance. It reflects hard-won lessons from regulated sectors where accountability is not optional.
Verkflöde publishes VAOM, updates it, and engages selectively on framework reviews, workshops, and speaking invitations.
Get in touch
For framework questions, feedback, speaking invitations, or collaboration.