Executives, senior leaders, and decision-makers responsible for approving AI initiatives, budgets, and organisational direction.
What it assesses Whether AI adoption is strategically grounded — including clarity of business objectives, prioritised use cases, ownership models, rollout discipline, and decision gates.
How it helps This questionnaire turns abstract “AI strategy” into explicit decisions. It surfaces whether leadership has defined what AI is for, what it should not be used for, who owns outcomes, and how risk is managed before scale. Outputs are designed to support board-level and senior leadership discussions.
Best used when
Deciding whether to scale AI beyond pilots
Aligning leadership on priorities and risk tolerance
This assessment helps leaders identify whether their organisation is ready to adopt AI without chaos. It focuses on strategy, ownership, governance, and rollout discipline.
Status:Not scoredCoverage:0%Score:—Decision:—
Your role (optional)
— Select role —
Executive / Strategy
Product / Operations
Engineering
Risk / Governance / Compliance
HR / Ethics
Other
Company or project identifier (optional)
Your score
0
out of 45
Answered 0/15
Posture
—
Answer the questions to see your posture.
Section A — Strategy & Value
1) Do you have a clear business objective for AI that is measurable (time saved, cost reduced, revenue increased, risk reduced)?
No — exploratory, not measurable
Partial — broad intent, weak metrics
Mostly — goals defined and measurable
Enforced — goals + metrics + named owners
2) Are your AI use cases prioritised (top 3–5) with an explicit “do not do” list?
No
Partial — informal prioritisation
Mostly — documented list exists
Enforced — scope boundaries are documented + used
3) Do you know what data the system will rely on (sources, freshness, ownership, access rights)?
No
Partial — some sources known
Mostly — sources mapped + owners known
Enforced — sources + freshness + rights governed
4) Do you have a “safe to ship” definition for AI outputs (quality threshold, refusal behaviour, escalation for high-stakes)?
No
Partial — informal expectations
Mostly — documented thresholds exist
Enforced — documented + tested + monitored gates
5) Do you have a defined adoption roadmap (pilot → gated production → scale) with decision gates?
No
Partial — pilots only
Mostly — roadmap exists
Enforced — roadmap + gates + accountability
Section B — Ownership & Operating Model
6) Are owners assigned for business value, product delivery, engineering reliability, and governance?
No
Partial — some owners, unclear gaps
Mostly — most owners assigned
Enforced — fully defined + active cadence
7) Do you have an intake process for new AI use cases (risk tiering, approval, prioritisation)?
No
Partial — case-by-case
Mostly — defined process
Enforced — defined + used + logged
8) Do you have a policy for which tasks require a human decision (human-in-the-loop) vs can be automated?
No
Partial — informal norms
Mostly — documented policy
Enforced — documented + enforced + audited
9) Do you have training and enablement (what to trust, when to escalate, how to report failures)?
No
Partial — ad-hoc training
Mostly — materials exist
Enforced — training + tracking + refresh cycle
10) Are vendors/tools evaluated with evidence requirements (logs, exportability, failure cases) rather than demos only?
No
Partial — some evaluation
Mostly — consistent criteria
Enforced — failure-path demos + sign-off
Section C — Governance, Risk & Continuity
11) Do you have incident response for AI failures (triage, rollback, comms, learning loop)?
No
Partial — informal
Mostly — documented
Enforced — practiced + measurable
12) Do you track drift over time (quality shifts, refusal changes, source shifts, new failure patterns)?
No
Partial — occasionally
Mostly — regular monitoring
Enforced — monitoring + alerting + cadence
13) Do you maintain change logs and approvals for prompts, models, tools, and data access?
No
Partial — ad-hoc
Mostly — consistent
Enforced — consistent + audited
14) Can you produce an evidence trail for outputs (sources used, checks applied, why it answered)?
No
Partial — limited traces
Mostly — good traces
Enforced — fully exportable evidence trail
15) Do you run structured evaluations (not demos) before shipping changes and at regular intervals?
