01 · Read
Which information enters the task?
List documents, databases, messages, sensors, and derived context. Include sensitive or stale information that may alter a recommendation even when the workflow cannot edit it.
Foundation 02 · Resource scope
Map what the
workflow can touch.
Before an autonomous workflow can receive a right, turn its description into a resource inventory. A task may read, alter, disclose, reserve, recommend, or commit more than its interface makes visible.
Learning boundary. This is an operating framework, not a safety certification, a legal authority model, or approval to grant a system access to data, funds, services, or people.
Learning objective
Turn “the agent handles it” into a testable inventory.
By the end of this lesson, you should be able to name the resource types a workflow may affect, separate direct access from downstream commitments, and identify the owner who must set the limit.
02 · Write
Which records can change?
Separate drafts from writes. A change to a ticket, calendar, inventory record, customer profile, or public page creates different reversal and review requirements.
03 · Commit
Which downstream obligations can begin?
Identify spend, reservations, notifications, work orders, vendor requests, and actions that consume another team’s time or change a person’s available choices.
04 · Depend
What breaks or changes if the input is wrong?
Record dependencies, the cost of delay, and the safe fallback state. A narrow tool call may still create a broad effect through automation around it.
Practice lab
Inventory a maintenance triage workflow.
A workflow reads equipment alerts and maintenance history, ranks candidate work orders, can update a ticket priority, and may draft a vendor message. A supervisor may later approve an outage window.
- Write one resource entry for every read, write, commitment, and dependency.
- Mark which entry could affect cost, schedule, safety, privacy, or another team’s rights.
- Name the owner and maximum allowed effect for the three highest-consequence entries.
- State the safe fallback when confidence, data freshness, or scope is uncertain.
Source shelf
Inspect the framework behind the exercise.
- NIST AI Risk Management FrameworkA voluntary framework for incorporating trustworthiness considerations into the design, development, use, and evaluation of AI systems.
- Autonomous Resource Management referenceUse this pathway’s framework reference as a starting point; do not infer public availability, implementation, or performance from the lesson.
Continue the pathway