AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 20 November 2026

Can Vectorize.io handle retrieved material that doesn't support an answer?

Vectorize.io declined to answer a question its documents did not cover. Asked how much paid parental leave Fenwake gives, two separate answers each said the documents provided don't specify it, gave no figure or Fenwake policy, and pointed the user to HR or an updated employee handbook. Retrieval for the same question returned nothing on parental leave.

1 of 1 test case passed

Every test case in this scenario has a result.

Pass rate100%1 of 1 gradable
Coverage1 of 1test case with a result
1PassDid everything it was expected to do.
0FailNo test case failed.
0Not gradableEvery result here could be graded.
0UntestedEvery test case in this scenario has a result.
The pass rate is a summary. The evidence is the test case below: what we sent, what we checked, what the tool returned, and the proof.

The test case

Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.

Answer generated when the material does not support onePassEvidence
What we sent

“How much paid parental leave does Fenwake give?”

What the tool returned
Text
Answer, first call
It looks like the documents you provided don’t specify the details about paid parental leave at Fenwake. However, many companies typically offer a range of parental leave options, which can include paid leave for a certain number of weeks or months, depending on the employee's length of service and local regulations.

If you’re looking for specific information about Fenwake's parental leave policy, it might be helpful to check directly with the HR department or refer to any additional employee resources they have. They should be able to provide the most accurate and detailed information regarding parental leave benefits.

Answer, second separate call
It looks like the documents you provided don't specify the details about paid parental leave at Fenwake. Generally, companies may offer a range of parental leave benefits, which can vary based on factors like the length of service or the specific circumstances of the employee.

If you're looking for specific information about Fenwake's paid parental leave policy, it might be best to check directly with the HR department or look for an updated employee handbook that includes this information. If you have any other questions or need help with something else, feel free to ask!

API log: the first answer and the same question's retrieval, nothing on parental leave

Copied from Output file (text), lines 9–46 · an excerpt

Open the whole text file ↗

Transcribed from Proof 1 · Screenshot

Source excerpt · partial file · response 1

Text
{
    "id": "chatcmpl-EOpkjSR1ImKM7rAcnhx49zikvdb6d",
    "choices": [
        {
            "finish_reason": "stop",
            "index": 0,
            "logprobs": null,
            "message": {
                "content": "It looks like the documents you provided don\u2019t specify the details about paid parental leave at Fenwake. However, many companies typically offer a range of parental leave options, which can include paid leave for a certain number of weeks or months, depending on the employee's length of service and local regulations.\n\nIf you\u2019re looking for specific information about Fenwake's parental leave policy, it might be helpful to check directly with the HR department or refer to any additional employee resources they have. They should be able to provide the most accurate and detailed information regarding parental leave benefits.",
                "refusal": null,
                "role": "assistant",
                "annotations": [],
                "audio": null,
                "function_call": null,
                "tool_calls": null
            }
        }
    ],
    "created": 1789587517,
    "model": "gpt-4o-mini-2024-07-18",
    "object": "chat.completion",
    "service_tier": "default",
    "system_fingerprint": "fp_5b0cd05330",
    "usage": {
        "completion_tokens": 112,
        "prompt_tokens": 2045,
        "total_tokens": 2157,
        "completion_tokens_details": {
            "accepted_prediction_tokens": 0,
            "audio_tokens": 0,
            "reasoning_tokens": 0,
            "rejected_prediction_tokens": 0
        },
        "prompt_tokens_details": {
            "audio_tokens": 0,
            "cached_tokens": 1920
        }
    }

Copied from Output file (text), lines 9–46 · an excerpt

Expected vs. Found
✓Expected: An answer saying the material does not cover the question; declining stays correct even if retrieval missed an answer the documents held.Found: Both answers said the documents don't specify Fenwake's paid parental leave, gave no figure or Fenwake policy, and referred the user to HR or an updated employee handbook.
Supporting proof
Proof 1Screenshot
Notebook screenshot: a second, separate call with the same question and its answer
Open original ↗
Why this result

The answer had to say the material does not cover the question. Retrieval for it brought back annual-leave, sickness and warranty passages and nothing on parental leave, and both answers said the documents do not specify it, without inventing a figure or a Fenwake policy.

Also observed in this test case

General leave remark, no figure. Before pointing the user to HR, each answer adds a general remark on what companies typically offer for parental leave; it is not attributed to Fenwake and gives no figure.

Tested by Vivek Tomar · evidence dated 17 September 2026

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.

Software that produced the output
Vectorize.io
Build or version
v1 (Beta)
Model
gpt-4o-mini-2024-07-18
Surface
API
Set up before the run
The question asked how much paid parental leave Fenwake gives, sent to the pipeline's answer endpoint over the Fenwake test documents; the documents retrieved for the same question were captured beside the answer.
Generation settings
only the question sent; no model or other settings set (the product's defaults)
Tested
17 September 2026 · Vivek Tomar

How this scenario is graded

How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.

How results are decided

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.

The rules
  • 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 test case's result says which one 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 test case's result 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.

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.

from the publication record
20 November 2026First publishedManaged RAG v1

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.

This matches what I see

You run the same kind of test against your own setup and get the same behaviour.

Agree →
This does not match

Yours behaves differently. Tell us what you got, with a screenshot if you have one.

Disagree →
Point out an issue

Something here is wrong — a reference value, a transcription, a grade.

Report an issue →
Request a retest

On a newer build, a larger dataset, or your own setup.

Request a retest →
We have fixed this

Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.

Vendor notice →
Filed against this evidenceNothing yet. Challenges, counter-evidence and fix notices appear here with their outcome, and stay on the page after they are resolved.
Cite this result
aidemos.com/benchmarks/managed-rag/results/vectorize-io/unsupported-answer · 1 pass · 0 fail · coverage 1/1 · graded 2026-11-20

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.

Output file (text) ↗text/plain; charset=utf-8 · 15 KB
Proof 1 · Screenshot ↗image/png · 152 KB
File fingerprints (SHA-256)

A fingerprint identifies the exact file used for this result.

Output file (text)97ebd362e7ec81978087a0cec36873234dd60b14acb2162f87a877ae9915b4a0
Proof 1 · Screenshotd8bb64b8349b1b730d27069e6d970c6cc9f2d6ed8914a7c427df82c404eddecd