Specifying the Delegated-Autonomy Boundary: Requirements Engineering for Agentic AI
Chetan Arora, Andreas Vogelsang, Abbi Sharma
Why It Matters
What makes this one worth your time
Understanding and specifying the delegated-autonomy boundary is crucial for developing safe and effective AI systems that operate with varying levels of autonomy.
The paper proposes a framework for defining and managing the delegated-autonomy boundary in agentic AI systems.
Summary
The paper introduces the concept of the delegated-autonomy boundary in agentic AI systems, proposing two artifacts: an Agency Justification Record (AJR) and an Agentic Delegation Policy (ADP) to guide the development of AI systems with varying degrees of autonomy.
Key contributions
- Introduction of the delegated-autonomy boundary concept.
- Proposal of the Agency Justification Record (AJR) for evaluating the need for agentic AI.
- Development of the Agentic Delegation Policy (ADP) for specifying autonomy-related requirements.
Notable insights
- The concept of graduated authority in AI systems, allowing for tiered control structures.
- The introduction of AJR and ADP as tools for requirements engineering in agentic AI.
Possible limitations
- Not stated in the abstract
Abstract
arXiv:2607.17225v1 Announce Type: cross Abstract: Agentic AI systems do not just predict or recommend; they plan, maintain state, and act in external environments with varying degrees of autonomy. This changes the requirements engineering problem in a specific and under-addressed way: it introduces what we call the delegated-autonomy boundary -- the set of decisions about what may be delegated to the system, under what graduated authority, with what oversight, and how control is returned. Current practices bury these decisions inside prompts, tool schemas, and runtime policies, even though they are requirements-level commitments. This paper proposes two complementary artifacts. First, an Agency Justification Record (AJR) helps teams decide when an agent is warranted over simpler alternatives. Second, an Agentic Delegation Policy (ADP) captures what must be specified for safe and effective development: purpose, authority, information, coordination, assurance, and evolution. Crucially, authority in the ADP is modelled as graduated, i.e., a tiered structure. We illustrate the framework with two contrasting examples: a safety-critical hospital discharge coordination agent and an automated code review agent.