Which record types should I request for an approval-authority task pack?
A checklist of record types that make a task pack complete enough to grade authority-dependent decisions, from limits and delegations to gold outcomes.
Use this checklist when drafting a request, so every task can be graded from the records supplied.
- **Authority matrix:** roles, limit types, values, measurement basis and effective dates. - **Delegation and acting records:** grantor, grantee, scope, start, end and any revocation event. - **Request records:** requester role, amount, category, risk flag, supporting evidence and decision date. - **Routing history:** prior approvals, returns, comments and escalations with timestamps. - **Policy text:** written rules that agree with the structured limits, plus any intended ambiguities. - **Gold outcomes:** permitted decisions per task, valid approver roles and rationale tags. - **Condition tags:** labels such as boundary-amount, expired-grant or linked-requests for per-condition reporting. - **Organisation structure:** reporting lines at role level so escalation targets can be resolved.
Not every pack needs every item. A question-and-answer task pack may fold delegation facts into the prompt, whereas a simulated environment stores them in tables the agent queries through tools.
The records are generated to fit the brief. Ask how the sample was grounded, check that the policy text and the structured limits agree, and review the output for disclosure risk before deciding on a larger scope.
Requesting the right record types up front avoids tasks that cannot be graded because a limit, date or gold outcome is missing.
- Audience:
- Evaluation teams
- Content type:
- Checklist
- Buyer question:
- What can be generated
Content release: r-a220eb6e099b243b
Related questions
- What fields and links make up a generated approval record schema?
- Can the generated authority matrix follow the structure of our own approval process?
- How can a task pack represent a claim settlement split to stay under a limit?
- Which records can a generated approval dataset contain?
- What should I specify when requesting generated authority and approval data?
- What does a generated claims settlement approval case look like?