Can FutureSmart Document Intelligence tell when the knowledge base has no answer?
FutureSmart Document Intelligence fails this scenario on the /search surface. The unanswerable paid-parental-leave query returns three sections with a top score of 1.0000, and the returned handbook sections do not answer it. The control query on the same corpus also returns three sections with a top score of 1.0000, and its top section contains the answer. A separate /chat answer says the handbook does not specify any paid parental leave entitlement.
1 of 1 test case failed
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 is based on this test case.
Reasonable question no document answersFailEvidence
“How much paid parental leave does Fenwake give?”
--- API call --- POST /collections/{CID}/search Body: {"query": "How much paid parental leave does Fenwake give?"} (default params — no top_k, no filters) --- Raw retrieved response --- status: 200 | sections returned: 3 score=1.0000 document_id=09c99644-e630-44a4-81f5-636ed80d942c Extracted Text: Employee Handbook Fenwake Instruments Ltd - all locations 1. Working here This handbook sets out the terms that apply to everyone employed by Fenwake Instruments Ltd, whether based at the Bristol workshop, at the Avonmouth calibration laboratory, or working from a customer site or vessel. It forms part of your contract. Where a term is set out in your written statement of particulars and also here, the written statement governs. 2. Hours of work Standard hours are 37.5 a week, worked between 08:00 and 18:00, Monday to Friday. Workshop staff work a 07:30 start by agreement with the workshop manager. Anyone rostered to the out-of-hours duty engineer rota is paid a standby allowance for each week on the rota, and an hourly rate for any hour actually worked while on standby. 2.1 Time recording Hours are recorded against a job number on the day they are worked. A job number is required for chargeable work, for internal projects and for training. Corrections are made by the line manager rather than by editing a submitted week. 3. Annual leave Full-time employees accrue 26 days of annual leave in each complete leave year, which runs from 1 April to 31 March. Part-time employees accrue pro rata. Leave accrues from the first day of employment and may be taken during the probation period with the line manager's agreement. 3.1 Booking leave Annual leave is requested in the rota system and confirmed by the line manager. Requests are considered in the order they are received. Service operations will decline a request that would leave the duty engineer rota uncovered, and will say so within five working days so that alternative dates can be chosen. 3.2 Carry-over Up to five days of untaken annual leave may be carried into the next leave year, and must be taken by 30 June. Anything beyond five days is carried over only where leave could not be taken because of a business requirement, and only with the written agreement of a director. 3.3 Leave on leaving score=0.5333 document_id=09c99644-e630-44a4-81f5-636ed80d942c 3.3 Leave on leaving On termination, accrued and untaken annual leave is paid in the final salary payment. Where more leave has been taken than accrued, the balance is deducted from that payment. 4. Sickness absence Report sickness to your line manager before 09:00 on the first day, by telephone rather than by message. Absence of seven calendar days or fewer is self-certified on return. Longer absence requires a fit note. Company sick pay is paid at normal salary for the first four weeks of a continuous absence and at half salary for the following four weeks, after which statutory sick pay alone applies. 4.1 Return to work Every absence is followed by a return-to-work conversation with the line manager. The conversation is recorded in one paragraph on the personnel file. Where an absence relates to a condition that affects offshore or diving work, occupational health is consulted before the employee returns to that work. 5. Expenses Expenses are claimed monthly, with receipts, and reimbursed with the following month's salary. Mileage in a private vehicle is paid at the approved rate. Overnight accommodation is booked through the company travel account wherever possible; where it is not, a room is reimbursed at cost within the regional cap published on the intranet. 6. Notice During probation, notice is one week on either side. After probation, notice is one month for most roles and three months for directors and for the head of service operations. The company may pay in lieu of notice. 7. Training and certification Offshore survival and first-aid certification is renewed at company expense before it expires. Time spent in required training counts as working time. Where an employee leaves within twelve months of a discretionary course costing more than two thousand pounds, a proportion is recoverable, as set out in the training agreement signed before the course. 8. Conduct at sea and on customer sites score=0.4167 document_id=09c99644-e630-44a4-81f5-636ed80d942c 8. Conduct at sea and on customer sites On a vessel, the master's instructions take precedence over any instruction from Fenwake. On a customer site, the site's permit-to-work system governs. An employee who believes work cannot be done safely is expected to stop it, and will be supported for having done so. 9. Personal data Personnel records are held for the duration of employment and for six years afterwards. Employees may ask to see their record at any time by writing to the operations director. --- Control: TC90's answerable question on the same corpus (score comparison) --- POST /collections/{CID}/search Body: {"query": "How quickly does Fenwake acknowledge a Priority 1 fault reported outside business hours?"} status: 200 | sections returned: 3 score=1.0000 document_id=88399604-1153-40ef-be66-f4da43a873a2 Where a site requires a permit, an induction or a vessel movement, the attendance clock is suspended until access is confirmed in writing. The suspension is recorded on the ticket with the reason. Fenwake internal - service operations Page 3 --- Fenwake Instruments Ltd - Field Service Policy and Charges (United Kingdom) # 6. **Out-of-hours cover** Outside business hours the service desk is not staffed. A duty engineer rota covers every hour of the year, and the out-of-hours number in section 8 pages the engineer on call. ## 6.1 **Acknowledgement out of hours** A Priority 1 fault reported outside business hours is acknowledged by a named duty engineer within 90 minutes of the ticket being raised. The engineer confirms ownership by telephone and, where the customer asks for it, in writing on the ticket. The target applies on every cover level; the cover level determines what happens after acknowledgement, not how quickly the call is answered. ## 6.2 **What follows acknowledgement** Having acknowledged, the duty engineer attempts a remote diagnosis using the telemetry record. Where the fault cannot be cleared remotely, attendance is arranged under section 5 from the start of the following working day, except on Critical cover, where the duty engineer may authorise attendance immediately. ## 6.3 **Priority 2 and Priority 3 out of hours** Priority 2 and Priority 3 tickets raised outside business hours are queued and acknowledged when the service desk opens. The business-hours targets in section 4 then apply from opening. Fenwake internal - service operations Page 4 --- Fenwake Instruments Ltd - Field Service Policy and Charges (United Kingdom) # 7. **Escalation** A customer may escalate at any time, and the service desk escalates on its own initiative when a target is missed. Each step below is a named person, not a queue. score=0.5833 document_id=88399604-1153-40ef-be66-f4da43a873a2 File Name: A01_field_service_policy_and_charges_uk.pdf Extracted Text: Fenwake Instruments Ltd # Field Service Policy and Charges United Kingdom operations ## 1. Purpose This document states what a customer holding a support contract with Fenwake Instruments Ltd may expect when an instrument fails in service, what the service desk in Bristol is obliged to do in response, and what is charged for work that falls outside cover. It applies to the Tolvan A3, the Tolvan C5 and the Brannoc telemetry gateway, on every cover level. ## 2. Scope Part A covers fault handling: acknowledgement, attendance, spares and escalation. Part B sets out the charges that apply outside cover. Neither part describes the goods warranty carried by a new instrument. ## 3. Fault priorities Every ticket is given a priority when it is raised. The priority is set by the effect on the customer's operation, not by which component failed. - **Priority 1** - a deployed instrument is not returning data, or is returning data known to be invalid, and no spare is rigged. - **Priority 2** - an instrument is degraded but usable, or a spare has been rigged and the primary unit is down. - **Priority 3** - a question, a configuration change, or a fault with no operational effect. A priority may be raised by the duty engineer at any point. It is lowered only with the customer's agreement, recorded on the ticket. Fenwake internal - service operations Page 1 --- Fenwake Instruments Ltd - Field Service Policy and Charges (United Kingdom) # 4. Acknowledgement during business hours Business hours are 08:00 to 18:00, Monday to Friday, excluding statutory non-working days in England and Wales. Within those hours the service desk is staffed continuously and the targets below are measured from the moment the ticket is raised, by telephone or through the portal, whichever is earlier. score=0.5333 document_id=88399604-1153-40ef-be66-f4da43a873a2 <table> <tr> <th>Priority</th> <th>Acknowledgement target</th> <th>Measured from</th> </tr> <tr> <td>Priority 1</td> <td>30 minutes</td> <td>Ticket raised</td> </tr> <tr> <td>Priority 2</td> <td>4 working hours</td> <td>Ticket raised</td> </tr> <tr> <td>Priority 3</td> <td>The next working day</td> <td>Ticket raised</td> </tr> </table> An acknowledgement is a named engineer taking ownership of the ticket and telling the customer what happens next. An automatic reply from the ticketing system is not an acknowledgement and does not stop the clock. ## 4.1 What the target does not cover The acknowledgement target is not a repair target. Attendance and parts are dealt with in section 5, and the two clocks run independently: a ticket may be acknowledged inside target and still take several days to close. Fenwake internal - service operations Page 2 --- Fenwake Instruments Ltd - Field Service Policy and Charges (United Kingdom) # 5. Attendance and parts Where a fault cannot be cleared remotely, the service desk arranges attendance. Attendance targets are working-day targets and are quoted from the point at which the customer confirms site access. <table> <tr> <th>Cover level</th> <th>Attendance target</th> <th>Spares held</th> </tr> <tr> <td>Standard</td> <td>Five working days</td> <td>Central store only</td> </tr> <tr> <td>Extended</td> <td>Three working days</td> <td>Central store only</td> </tr> <tr> <td>Critical</td> <td>Next working day</td> <td>Regional depot</td> </tr> </table> ## 5.1 Consumables Anti-fouling guards, sacrificial anodes and cable glands are consumables. They are replaced during any attendance and charged at the rates in Part B. ## 5.2 Access Where a site requires a permit, an induction or a vessel movement, the attendance clock is suspended until access is confirmed in writing. The suspension is recorded on the ticket with the reason.
Transcript of the search and the control question, with the returned sections shown in full.
Copied from Captured file (text), lines 1–7, 9–135
Open the text file ↗| ✗ | Expected: It should clearly show no match, such as no results or a clear score gap.Found: It returned three sections with a top score of 1.0000, matching the answerable query's shape. |
The /search response was checked against the unanswerable question and the control query. The unanswerable query returned three sections with a top score of 1.0000, but the returned handbook sections do not answer it. The control query returned the answer in its top section.
Chat answer shown. The chat answer says the handbook does not specify any paid parental leave entitlement.
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 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.
| Level | Name | Scope |
|---|---|---|
| Benchmark | Managed RAG → | 15 scenarios · 7 tools |
| Capability | Retrieval → | |
| Scenario | Answer is not available in the knowledge base → | |
| 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.