Skip to content

Industry field guide

Computer facilities restore and recovery evidence for AI

A managed facility can see a backup marked complete before anyone has verified a restore.

2022 NAICS
541513
Updated

What this guide tests

Which system has a verified restore test, and can the backup labels or Test-A inventory establish production recovery for either system?

See the specimen and fields · Work through the fictional records

Task and source context

A managed facility can see a backup marked complete before anyone has verified a restore. NIST’s contingency planning guide includes recovery strategies and testing. A facilities operator can build an offline evidence exercise using dummy systems and made-up backup records. NIST contingency planning guide

Create a mock backup event, a restoration job for a test target and a report that verifies a dummy file inventory. Add a completed backup for a different system with no restore evidence. Ask the model which system has a verified test, what was actually checked and whether the result applies to production. Score role, system and result independently.

Use original exports or mock tables. Keep topology, customer information, passwords, recovery locations and live controls out. Review client and operator rights before considering real logs.

Describe a fictional restore verification example

Industry research

The U.S. Census Bureau’s industry definitions establish which activities these codes cover. The workflow sources in this guide describe the task or record type. Fiund’s pilot below uses fictional material to test a specific decision within that scope.

541513 · Computer Facilities Management Services

This U.S. industry comprises establishments primarily engaged in providing on-site management and operation of clients' computer systems and/or data processing facilities. Establishments providing computer systems or data processing facilities support services are included in this industry.

U.S. Census Bureau, 2022 NAICS Manual, p. 465 ↗

Build a pilot sample

Set up this pilot

Build two dummy restore logs with equivalent backup labels and different test targets. Require the inventory and reviewer card for each target and keep production recovery unverified.

Prepare the test packets

  1. Start with the worked case. Use the fictional record types below as a guide to the source files you need. Give each packet a stable ID and save its source versions with it.
  2. Change one fact. Make one packet with enough evidence to answer the question and one where the evidence points to a different answer. Keep the other fields fixed so a reviewer can tell what changed.
  3. Remove the decisive evidence. Give the reviewer a third packet with a plausible near-match in its place. Have them write what can still be established and what evidence would resolve the rest.

Fields the reviewer needs

System, backup and test-target IDs
Connects each stage to the correct artifact.
Verified inventory and review source
Defines what the test actually checked.
Production versus test scope
Prevents a lab result becoming deployment authority.

Use Build an AI evaluation from professional work to turn the specimen into a brief with a defined task, source and review owner.

Fictional exercise

A completed backup and a verified restore test prove different stages

A detached facilities exercise uses dummy systems and a made-up file inventory. No network, topology, credential, recovery location or live control is present.

Records supplied to the evaluator
Record IDSourceWhat it records
BR1Practice backup logSystem A backup B-A completed. System B backup B-B completed separately.
BR2Test-target restoreB-A restored to isolated target Test-A; independent inventory check confirms three named dummy files.
BR3Scope statementTest-A check covers dummy-file presence only. No application behavior or production recovery approval recorded; B has no restore test.

Question for the model. Which system has a verified restore test, and can the backup labels or Test-A inventory establish production recovery for either system?

Show the reference answer

BR2 verifies a restore of A's B-A backup to Test-A and the presence of the three dummy files. BR3 limits that result to the stated inventory check, without application or production-recovery evidence. B's completed backup in BR1 has no restore test at all. Report backup completion, test restoration, checked scope and absent approvals separately rather than treating every completed backup as demonstrated recovery.

Download the fictional records and answer key (JSON)

How to check the result

Ask for the answer, the record IDs that support it and any fact the evidence leaves unresolved. Have a reviewer compare the response against the reference before deciding whether the sample is ready to expand.

Errors to flag in this case

  • Calling backup completion a successful restore test.
  • Extending a file-presence check to full application recovery or from system A to B.

Change the inputs and run it again

Add a B restore job marked completed without an independent file check. Require the model to distinguish job completion from verified contents.

Score the decision and the evidence separately

Save the reviewer’s expected answer, supporting record IDs and unresolved fields in a separate answer key. Score the model on the decision, its citations and its handling of the deliberate gap. A correct answer citing a different case counts as an evidence error.

Before expanding collection, have a second reviewer run the same checks on the held-out packet. Keep related versions, views and exports of the same case together so they stay on one side of the training and evaluation split.

Check permission and scope

Use original mock tables. Exclude customers, topology, passwords, recovery locations and live operating controls.

Use original or licensed practice material. Leave client work product, confidentiality obligations and regulated outcomes out of the sample.

Describe an authorized archive through Fiund’s existing-data assessment, or describe a new capture capability through the collection brief. The contributor agreement covers the terms for participation.

Sources and further reading

Resources for this workflow

How the examples work

Company and program reviews

Use these reviews to compare sourcing, capture or evaluation approaches. Confirm the applicable program, accepted sources and terms with the provider.

All company reviews

Related industry examples

← Professional and technical services