Industry field guide
Emergency relief kit staging and handoff data for AI
A relief organization can track supplies staged, dispatched and received without linking the exercise to any affected household.
- 2022 NAICS
- 624230
- Practice pack
- Example and answer key (JSON)
- Updated
What this guide tests
Reconcile the six I1 boxes by supported state, and can the I2 receipt resolve K4's hold?
See the specimen and fields · Work through the fictional records
Task and source context
A relief organization can track supplies staged, dispatched and received without linking the exercise to any affected household. The U.S. Fire Administration describes resource management during emergency response. A tabletop training set can test whether a model recognizes partial receipt and preserves missing evidence. U.S. Fire Administration resource management
Prepare six inert kit boxes and original labels. Create a dispatch for four and a receipt for three, with a documented hold for the fourth. Ask the model to reconcile totals by state and cite the record for the held item. Include a receipt under another incident identifier.
Use fictional events and no live distribution details, real locations, volunteers without permission or household records. Keep operational decisions with qualified staff.
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.
624230 · Emergency and Other Relief Services
This industry comprises establishments primarily engaged in providing food, shelter, clothing, medical relief, resettlement, and counseling to victims of domestic or international disasters or conflicts (e.g., wars).
Build a pilot sample
Set up this pilot
Stage six numbered mock aid boxes with a hold, two dispatches and one acknowledgment. Keep destination and incident identifiers visible, then reconcile held and in-transit quantities.
Prepare the test packets
- 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.
- 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.
- 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
- Incident and kit alias
- Defines inventory identity.
- Staging, actual dispatch and hold reference
- Separates movement states.
- Receipt target and remaining balance
- Supports reconciliation.
Use What health care and social services teams can contribute to AI to turn the specimen into a brief with a defined task, source and review owner.
Fictional exercise
Staged, dispatched, held and received kit counts describe different states
Six inert empty boxes and original labels form a closed-room tabletop fixture. No live incident, recipient, distribution route or operational decision is represented.
| Record ID | Source | What it records |
|---|---|---|
| RK1 | Staging register | Fictional incident I1: boxes K1–K6 staged, total 6. |
| RK2 | Dispatch and hold | K1–K4 assigned to dispatch D1. Documented hold H4 retains K4 at staging; actual outgoing D1 lists K1–K3 only. |
| RK3 | Receipts | I1/D1 receipt explicitly names K1–K3. A separate I2 receipt names a K4-like label, with no I1 or H4 relationship. |
Question for the model. Reconcile the six I1 boxes by supported state, and can the I2 receipt resolve K4's hold?
Show the reference answer
RK1 starts with six staged boxes. RK2's actual outgoing list and RK3's I1 receipt support K1–K3 as dispatched and received. K4 is documented held at staging, while K5 and K6 have no outgoing event and remain only staged in this fixture. RK3's I2 receipt belongs to another incident and cannot clear H4. Distinguish planned dispatch from actual movement and receipt rather than counting four received from the original assignment.
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
- Counting planned dispatch quantities as actual receipt.
- Using another incident's similar label to close a held-item exception.
Change the inputs and run it again
Add a later approved K4 dispatch with no receipt. Require its state to move from held to dispatched, while received count remains three.
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 owned empty props and consenting adults. Exclude live locations, households and identifiable volunteer records; qualified staff retain response decisions.
Use fabricated scenarios without beneficiary details, addresses or any individual eligibility determination.
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
Includes worksheet
What health care and social services teams can contribute to AI
Separate a staff training simulation from a governed clinical or care-data project.
Includes worksheet
Build an AI evaluation sample for service handoffs
Use fictional records to test whether a handoff reached a documented state.
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.
Compare the approach
Protege review for AI teams and data owners
Compare the company’s healthcare data-provider pathway separately from a fictional staff workflow.
Compare the approach
Sapien review: dataset quality review and expert checks
Compare an independent quality review process for an administrative or nonclinical evaluation.
Related industry examples
Industry guide · NAICS 624221
Temporary shelter capacity and reporting window data for AI
A capacity snapshot can become stale quickly and may include spaces unavailable for a given reporting window.
Industry guide · NAICS 624229
Community housing repair scope and completion data for AI
A volunteer repair organization can demonstrate how it links an approved task list to a completion photo.
Industry guide · NAICS 624210
Food pantry case counts and distribution inventory data for AI
A pantry inventory may count cases at receipt and individual units when supplies are packed.