Can FutureSmart Document Intelligence handle retrieved sources that disagree with each other?
FutureSmart Document Intelligence showed that the retrieved sources disagreed about the warranty period. It reported 24 months from delivery in the customer-facing sales pack and 36 months from delivery in the service-desk service pack, with both periods beginning at delivery.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above counts this row.
Two documents give different values for the same factPassEvidence
“What is the warranty period on a new Tolvan A3?”
answer: The retrieved warranty documents conflict: the customer-facing sales pack states **24 months from delivery**, while the service-desk service pack states **36 months from delivery**. Both specify that the period begins at delivery, not commissioning.
Plain-text evidence file shows the warranty question, answer, and 27 retrieved sections sorted by score.
Copied from Output file (text), lines 15–17
Open the text file ↗These are scenario-level criteria. Each test case's Expected and Found are listed separately.
- Whether the answer recognises that the retrieved sources conflict on the same fact.
- Whether the answer avoids presenting one source’s value as the fact.
- Whether the answer conveys that the sources disagree instead of choosing a side without support.
| ✓ | Found: It named both warranty values with their sources instead of one silent fact. |
The tool returned both conflicting warranty values and attributed each to its source instead of choosing one as the fact. The evidence also shows both relevant documents were retrieved, so no silent single-value answer was produced.
Score is normalized. The top hit scored 1.0000, so the score is rank-normalized rather than an absolute confidence measure.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the row says which and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the row says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Managed RAG → | 15 scenarios · 5 tools |
| Capability | Grounded Generation → | |
| Scenario | The retrieved sources disagree with each other → | |
| Tool | FutureSmart Document Intelligence → |
Global scenario definition → · Global capability definition →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.