From Available to Viable

The core argument

Tool availability is not execution viability. As AI agents move from answering questions to acting in real systems, governance has to evaluate whether a proposed action is authorized, evidenced, constrained, recoverable, and operationally safe before execution.

Why this matters

Many agent stacks answer the wrong governance question. They can show which tools exist, which APIs are callable, or which permissions appear valid. But an available tool call can still be the wrong action when evidence is weak, authority is unclear, constraints are binding, or recovery would be difficult.

The white paper explains the gap between what an agent can do and what it should be allowed to execute in context.

What the paper covers

• The shift from model behavior to execution authority.
• Why permissions and scalar scores are not enough for agentic workflows.
• How the PV-PP framework separates availability from viability.
• A practical execution viability layer for agent/tool systems.
• How the PV-PP Agent Auditor can surface hidden governance gaps before deployment.

Who this is for

This paper is written for people working on AI agents, enterprise AI governance, workflow automation, risk, compliance, audit, and operational control.

It is especially relevant where agents may update records, route cases, trigger workflows, touch financial processes, send communications, classify exceptions, or mark work complete.

About the PV-PP framework

The Productive Value–Productive Power (PV-PP) Framework is a decision and governance architecture for evaluating action viability under authority, evidence, constraint, recovery, memory, and execution-risk conditions.

The framework is being developed by Lance Amundsen as an independent research and development project focused on AI agent governance and non-scalar decision control.