Querying
Operates over the extracted records — filters, totals, answers questions — and is clear when an answer didn't come from them.
What this capability means
Querying means a tool can work over the structured records it has already extracted, across documents: it can filter them, total them, and answer questions from them. The way the tool exposes that ability is not what defines the capability. It should also make clear when an answer did not come from those extracted records. A query is judged against the extracted records the tool actually returned, not against the source documents.
Boundary: It does not judge whether the original extraction was correct; that is scored elsewhere.
Scenarios that test this capability
A scenario is a real-world situation used to test a capability.
Benchmarks that include this capability
A capability has global identity and may be used by more than one benchmark.
Structured Document Extraction
Which document extraction platform turns a user-defined schema into correct structured records — across layouts, scans and repeated sets — and is honest about what it could not find?