Industry field guide
Process service attempts and proof records for AI
A service company can build an evidence exercise around assignment, attempt and later proof.
- 2022 NAICS
- 541199
- Practice pack
- Example and answer key (JSON)
- Updated
What this guide tests
As of day 6, which assignment has a linked filed-proof entry, and what state is documented for the other?
See the specimen and fields · Work through the fictional records
Task and source context
A service company can build an evidence exercise around assignment, attempt and later proof. The Federal Rules of Civil Procedure include provisions for proving service under Rule 4. The proposed dataset evaluates whether software distinguishes the records presented in a mock file. U.S. Courts rules of civil procedure
Create two invented assignments with matching due dates. Provide an unsuccessful attempt for one and a dummy filed proof record for the other. Include an older address note under a similar case number. Ask the model to identify the documented state of each assignment and cite the record, without declaring that service is legally sufficient.
Use fabricated case names and inert form designs. Leave real court papers, witness details, addresses, evidence photos and service route information out.
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.
541199 · All Other Legal Services
This U.S. industry comprises establishments of legal practitioners (except offices of lawyers and attorneys, settlement offices, and title abstract offices). These establishments are primarily engaged in providing specialized legal or paralegal services.
Build a pilot sample
Set up this pilot
Use two fictional assignments with attempts on different dates. Attach a filed record to only one and grade as-of attempt and filing states separately.
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
- Assignment and mock case
- Provides the exact join between register and proof.
- Attempt date and outcome
- Describes the event actually logged.
- Proof entry and filing date
- Separates later documentation from an attempt.
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
An attempt log and a filed proof describe different record states
Two invented assignments use inert case IDs and original mock paperwork. The exercise concerns file status, with no real case, address, witness or legal sufficiency determination.
| Record ID | Source | What it records |
|---|---|---|
| PS1 | Assignment register | Assignment A and assignment B both have a day-8 due date. A refers to mock case X; B to Y. |
| PS2 | Attempt log | A: day 3, unsuccessful attempt recorded. B: day 4, attempt entry with proof status left blank. |
| PS3 | Proof register | Original dummy filed-proof entry P7 links to B/Y on day 5. An older address note links to X-old, not X. |
Question for the model. As of day 6, which assignment has a linked filed-proof entry, and what state is documented for the other?
Show the reference answer
PS3 links P7 specifically to B and mock case Y, so B has a documented filed-proof entry by day 6. PS2 supports an unsuccessful attempt for A; neither A's due date nor the older X-old note supplies a proof for A/X. Report these source states separately. A filed entry does not, within this exercise, establish that service satisfies any applicable legal rule.
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
- Converting an assignment due date into evidence of completed service.
- Attaching a proof or address note through a similar case name while ignoring the assignment and case IDs.
Change the inputs and run it again
Add an unfiled draft proof for A dated day 7. The answer should acknowledge the draft without changing the day-6 filed-proof finding.
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 fictitious forms. Exclude court papers, signatures, routes and personal details; leave service sufficiency to qualified review.
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
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
Invisible Technologies review for expert-led AI training
Compare expert calibration and reviewer agreement before commissioning a professional-work evaluation.
Compare the approach
Mercor review for expert AI training work
Read how an individual expert role differs from licensing an employer or client’s work product.
Related industry examples
Industry guide · NAICS 541191
Title and settlement fee lineage data for AI
A title office may receive an estimate, a revised invoice and a final settlement disclosure containing similar charge descriptions.
Industry guide · NAICS 541110
Law firm contract redline and approval evidence for AI
Legal AI researchers need tasks with a reviewable outcome.
Industry guide · NAICS 541612
HR consulting job analysis and competency evidence for AI
Job analysis connects tasks and required competencies.