Skip to content
MirrorFirm

Can generated tasks include conflicting or superseded correction policies?

Generated policy collections can include versions, effective dates and precedence rules that require agents to identify which instruction governs a particular adjustment.

**Can a task contain more than one plausible policy?**

Yes. A generated pack could include a current adjustment procedure, an earlier procedure retained for historical cases and a local team guide. Their status and authority should be encoded clearly enough that a careful agent can determine which applies.

**What makes the conflict useful?**

The conflict should affect a decision. An earlier procedure might permit team-level approval for a category that now requires central review. A local guide might describe how to prepare supporting evidence while lacking authority to change approval requirements. The agent must distinguish supplementary guidance from governing policy.

**What should the buyer inspect?**

Ask for policy identifiers, version relationships, effective periods and a stated precedence rule. Also ask whether applicability follows the original transaction date, the correction request date or another defined event. That choice should come from the supplied policy.

**Can ambiguity itself be generated?**

Yes, where the expected response is to identify an unresolved policy conflict and seek a ruling. The answer key should distinguish deliberate ambiguity from a missing rule that makes the task unanswerable.

Evaluate policy selection explicitly instead of assuming every document in the context has equal authority.

Audience:
Evaluation teams
Content type:
FAQ
Buyer question:
What can be generated

Content release: r-a220eb6e099b243b

More on Ledger adjustment