Assessment · Process

Assessment Process

A six-step process from exam authoring to verifiable credential — connecting authoring, grading, analytics, and delivery on one AI-competency data platform.

The process

Six steps, end to end.

  1. 01

    Author your exam

    In the Question Factory, a swarm of 5–10 authoring agents drafts your exam in parallel. You set the scope, item formats, difficulty, and pass criteria for each role, then refine the items with the agents through dialogue.

    Artifacts produced
    • Exam blueprint
    • Item bank (versioned)
    • Pass-criteria definition
  2. 02

    Cohort intake

    Your cohort roster is provisioned via SSO/SCIM. AIQCAT issues unique examinee credentials and confirms eligibility against the engagement agreement.

    Artifacts produced
    • Cohort roster (CSV / SCIM)
    • Eligibility confirmation
    • Examinee credentials
  3. 03

    Session & submission

    At the scheduled window the 120-minute session opens. Examinees produce deliverables in real formats — Excel, PDF, images, video. Files are encrypted on upload, timestamped, and version-tracked.

    Artifacts produced
    • Brief & rules
    • Isolated workspace
    • Submission receipt
  4. 04

    Consensus grading

    1,000 AI evaluator engines grade each submission against your rubric across all six dimensions. Scores are combined by consensus to reduce single-model bias; disagreement is flagged for human review.

    Artifacts produced
    • Rubric-mapped scores
    • Disagreement flags
    • Calibration audit log
  5. 05

    Examiner review

    A human examiner sample-reviews grading, resolves flags, and signs off the result. Borderline cases are escalated under the dispute-resolution protocol.

    Artifacts produced
    • Examiner sign-off
    • Sampled audit
    • Dispute log (if any)
  6. 06

    Report & credential

    A documented evaluation report — with red-pen annotations on the actual work — is delivered to the examinee and to your organization. Successful candidates receive a verifiable credential entered in the AIQCAT Credential Registry, and results roll up into your organization-wide dashboard.

    Artifacts produced
    • Evaluation report (PDF)
    • Red-pen annotations
    • Verifiable credential
Integrity

Documented, replayable, defensible.

Every grading decision can be replayed against the published rubric. Calibration runs against a held-out reference set; a sample of completed sessions is reviewed each cycle and summarized in the transparency report.