AI technical specification comparison

Compare stated requirements and supplier responses field by field. Preserve exceptions and unstated values instead of turning an incomplete response into a compliance claim.

Synthetic evaluation input

Requirement R1: deliver editable CSV register.
Requirement R2: include document revision register.
Supplier A: CSV register and revision register included.
Supplier B: PDF register only; revision register included.
Supplier C: CSV register included; no statement on revisions.
Allowed statuses: met; not_met; not_stated.

Input file Evaluation prompt Answer key

Original Inzonex training example, version 1.0. No customer document or real equipment data. No model has been scored on this page.

Expected fields

Field	Reference answer
A_R1	`"met"`
A_R2	`"met"`
B_R1	`"not_met"`
B_R2	`"met"`
C_R1	`"met"`
C_R2	`"not_stated"`

Check a structured response

The comparison runs in your browser against this answer key. It checks exact fields and values, not the accuracy of an explanation or a model's general ability.

Response JSON

Recommended workflow

Build a requirement-to-evidence matrix with separate met, not-met and not-stated states. Do not infer compliance from marketing descriptions.

A specification comparison needs the requirement, the supplier's statement and the reason for the assigned status. A missing statement is not the same as a failed requirement, and neither should silently become a pass.

The synthetic example uses document-delivery requirements rather than equipment design limits. One supplier satisfies both, another explicitly offers a different format, and a third says nothing about the revision register. The expected result distinguishes each case.

A real purchasing decision may depend on commercial qualifications, standards, project-specific definitions and signed deviations. Those documents must be included in the review. A comparison assistant can organise evidence but cannot invent supplier commitments or issue final technical acceptance.

Failure checks

Do not treat PDF-only delivery as editable CSV.
Distinguish not_stated from not_met.
Evaluate requirements separately for each supplier.

What this exercise does not prove

The exercise covers documentary comparison, not product qualification, procurement approval or standards compliance.

For an actual evaluation, keep a separate held-out set, record tool/model version and settings, and log raw outputs, corrections, elapsed time and actual charges. Do not compare tools tested on different inputs as though they ran the same benchmark.

What is checked in AI technical specification comparison?

Compare stated requirements and supplier responses field by field. Preserve exceptions and unstated values instead of turning an incomplete response into a compliance claim.

What does AI technical specification comparison not prove?

The exercise covers documentary comparison, not product qualification, procurement approval or standards compliance.

Methodology and reuse

These packets and answer keys are original Inzonex educational material, licensed CC BY 4.0. Attribute Inzonex and link to this task page when reusing the packet. The licence does not cover third-party material linked from this site.

NIST AI 600-1: background on generative AI evaluation and risk. This exercise is not NIST-certified.