YOURCoOperator track handout · 7 slides
Assessor ↗Skill suite ↗Back to deck
Shared · Opening01 / 07

Why this training exists

AI Use-Case Assessor — Workforce Training · Operator track

  • 01

    “Have you trained and informed the employees using AI?” is becoming an audit question.

  • 02

    This training and the assessor are how the organization answers yes.

  • 03

    Governance names the controls. This is where they reach the people who build and buy AI.

REF AI-ASS-SHARED-01SHARED · CONFIDENTIAL — INTERNAL

SHARED 1 · Narration

Every AI governance program eventually meets one question in an audit: can you show your people know what to do. Policy on a shelf does not answer it. This short training does. By the end you will know what counts as an AI use case, why we need to know about it, and how to run it through our assessor before it goes live.

Shared · Definition02 / 07

What is an AI use case?

  • 01

    Any system that uses AI or automated decision-making to predict, score, classify, generate, recommend, or decide.

  • 02

    If you are building, buying, or turning on a tool that does any of these, it is an AI use case — even if a vendor runs it.

  • 03

    Examples: scoring job applicants, a customer chatbot, flagging transactions, drafting documents.

REF AI-ASS-SHARED-02SHARED · CONFIDENTIAL — INTERNAL

SHARED 2 · Narration

The skill that matters most is recognizing one, and you do not need to be technical to have it. If a tool predicts, scores, classifies, generates, recommends, or decides, especially if it touches people, it is an AI use case we need to know about. Bought from a vendor still counts. When in doubt, submit it.

Shared · The tool03 / 07

What the assessor gives you

  • 01

    A preliminary risk tier, with the reasoning.

  • 02

    An explicit list of what your description did not say: the gaps.

  • 03

    The governance frameworks and obligations that likely apply.

  • 04

    It is advisory. A human reviewer confirms the outcome.

REF AI-ASS-SHARED-03SHARED · CONFIDENTIAL — INTERNAL

SHARED 3 · Narration

The assessor is not a gatekeeper to fear. You describe your use case and it returns a preliminary tier and why, the frameworks that likely apply, and, importantly, a plain list of what it could not tell from your description. That gap list is a to-do, not a failure. It tells you what to go find out.

Operator track · Practice04 / 07

How to run it

  • 01

    Install the suite in Claude Code (terminal).

  • 02

    Run the assessment, then paste a brief, README, filled form, or plain description.

  • 03

    Or use the same web assessor submitters use: https://www.mercedesedwards.com/assess

$/plugin marketplace add https://github.com/MercedesDG/ai-governance-suite.git

$/plugin install ai-usecase-assessor@ai-governance-suite

$/ai-usecase-assessment

REF AI-ASS-OPR-04OPERATOR TRACK · CONFIDENTIAL — INTERNAL

OPERATOR 4 · Narration

If you operate the tool, you can run it in the terminal. Add the marketplace, install the suite, and invoke the assessment, then paste any brief, README, or filled form. You get the same neutral result as the web form, with the full structured output.

Operator track · Outcome05 / 07

Understanding the output

  • 01

    Two stages: the parser extracts a neutral, tiered profile and names its gaps; the matcher maps it to frameworks and control objectives.

  • 02

    The tier is severity triage. The governance signals are a separate axis.

  • 03

    “Presumptive” means a framework's trigger (jurisdiction, AIMS) could not be resolved from the input — not that it does not apply.

  • 04

    The output feeds the system of record, where the record is kept and monitored.

REF AI-ASS-OPR-05OPERATOR TRACK · CONFIDENTIAL — INTERNAL

OPERATOR 5 · Narration

Under the hood, two stages run. A parser turns the description into a neutral, tiered profile and names its gaps, and a matcher maps that to the frameworks and control objectives that apply. When a framework comes back presumptive, it means the input could not confirm its trigger, like jurisdiction, not that it is irrelevant. This output feeds our system of record, where the record is kept and monitored.

Shared · Accountability06 / 07

Your responsibility

  • 01

    Submit before you build, buy, or turn it on — not after.

  • 02

    Keep it honest. Flag what you do not know.

  • 03

    Re-assess when the use case materially changes.

  • 04

    The assessment starts the record; it does not end your accountability.

REF AI-ASS-SHARED-06SHARED · CONFIDENTIAL — INTERNAL

SHARED 6 · Narration

Two rules. Submit early, before the thing is live, when governance can still shape it. And keep it honest, because the tool and the reviewer both rely on what you tell them. If the use case changes materially later, run it again.

Shared · Close07 / 07

Why this matters

  • 01

    Using the assessor and completing this training is how the organization shows its AI is governed — to a regulator, an auditor, or a customer.

  • 02

    Trained employees plus a record of every use case equals a defensible program.

  • 03

    Governance named the controls. You are how they operate.

REF AI-ASS-SHARED-07SHARED · CONFIDENTIAL — INTERNAL

SHARED 7 · Narration

When an auditor asks whether our people are trained and whether we know what AI we run, this training and this tool are the answer. You completing it, and submitting your use cases, is what turns our governance from a document into something we can stand behind. That is the whole point.