Workshop contents

Session 1 · Layer 5

Guardrails and Human Authority

What it may decide, and who signs

AI capability is no longer the constraint. The question is now: under what conditions, with what evidence, and with whose authority.

Step 5.1 — What could go wrong?

ScenarioWhat went wrong
AI recommends cancelling all First Class shippingNo cost or contract constraint applied, so it breaks customer commitments
AI sends the alert to the wrong executiveNo access control, so information leaks
AI bases a recommendation on corrupted dataNo data validation, so garbage in, garbage out
AI recommendation violates a vendor contractNo legal or compliance context, so liability
AI optimises for speed and ignores carbonNo sustainability guardrail, so misaligned KPIs
Facilitator asks the room

Which of these has already happened in your organisation — or could happen tomorrow?

Step 5.2 — The governance layer

Guardrails for supply chain AIWorkshop ScenarioClick the text, then ⌘C or Ctrl+C
DECISION AUTHORITY:
- AI MAY: Generate reports, send alerts, draft recommendations
- AI MAY (with approval): Adjust shipping mode allocation up to 10%
- AI MAY NOT: Change carrier contracts, override SLA commitments,
  communicate to external partners

DATA RULES:
- Must use data less than 24 hours old
- Must flag if any data source is missing or incomplete
- Must show confidence level for predictions

EVIDENCE REQUIREMENTS:
- Every recommendation must cite specific data points
- Must show "what if" comparison (current vs recommended)
- Must include risk assessment for each action

ESCALATION:
- If late delivery rate exceeds 60%: escalate to COO immediately
- If recommended action exceeds budget by >5%: requires CFO approval
- If pattern is unprecedented: flag for human review, do not auto-act

AUDIT:
- Log every decision, recommendation, and data source
- Weekly governance review of all AI-generated actions

Lesson. Guardrails define what AI may do, under what conditions, and with whose authority. Without them, no AI deployment is trustworthy enough to operate.

Step 5.3 — Executive self-assessment: the AI Readiness Rubric

Each participant evaluates one of their own AI initiatives against six dimensions.

DimensionLevel 1 · Ad hocLevel 2 · StructuredLevel 3 · Enterprise
InstructionsVague, generic promptsSpecific, formatted, role-basedConstraints, reasoning direction, output standards
ContextNone, general knowledge onlyDataset providedBusiness rules, history, KPIs, constraints
ToolsChat onlyData analysisMulti-tool workflow: query, visualise, draft, send
OperationHuman runs every requestScheduled tasksAgent monitors and acts within defined authority
GovernanceNoneBasic access controlDecision authority, evidence rules, escalation, audit
Evidence“AI said so”Data-backedCited, compared, confidence-scored, peer-reviewed

Facilitator exercise, 3 minutes. Pick one AI initiative you are responsible for. Where does it sit on this rubric today? Where does it need to be in six months — and what is the first gap to close?

How to read the figures on this page Source Dataread straight from the DataCo dataset Derived Evidencecalculated from those values Workshop Scenarioinvented for the exercise; not in the dataset