How much AI is running in your organisation that nobody's governing?
Shadow AI — undisclosed tools, unsanctioned models, ungoverned decisions — is already inside your organisation. The employees using it don't think of themselves as breaking rules. The people designated to oversee your AI systems often don't know half of it exists.
This diagnostic tells you how exposed your oversight structure is, in 5 minutes.
3–5×
More AI tools in active use than most governance registers capture
73%
Of EU AI Act Article 14 oversight failures trace to capability gaps, not missing policy
Aug 2026
Full EU AI Act enforcement. Human oversight capability must be demonstrable — not assumed
What this is: A directional exposure signal based on your responses. It is not a formal AARI assessment — that requires a minimum of 5 respondents and produces audit-grade output. This diagnostic tells you where to look. The AARI Index tells you what is actually there.
Questions 1–2 of 8 · Shadow AI Inventory
What does your AI inventory actually capture?
Most governance registers reflect what IT approved. They rarely reflect what employees are using.
Q1 How many AI tools or systems are formally registered in your organisation's AI system inventory?
Q2 How confident are you that your registered inventory captures everything employees are actually using day-to-day?
Questions 3–4 of 8 · Oversight Coverage
Who is actually responsible for oversight — and is there a backup?
Article 14 requires designated, capable overseers. Single points of failure are a common and costly gap.
Q3 For your highest-risk AI system, how many people are formally designated as human overseers?
Q4 If your primary overseer was absent today, could another person perform meaningful oversight — or would the system effectively run ungoverned?
Questions 5–6 of 8 · Capability Readiness
Can your oversight team demonstrate what oversight actually means?
AI literacy training and oversight capability are not the same thing. Regulators are beginning to understand the difference.
Q5 In the last 12 months, has your oversight team received specific training on what meaningful human oversight of AI requires — not general AI literacy, but oversight?
Q6 If a regulator asked your AI overseers to demonstrate their oversight process today, how would you honestly rate your team's readiness?
Questions 7–8 of 8 · Leadership & Governance Culture
Is leadership genuinely driving governance — or just sponsoring it?
The gap between stated leadership commitment and active governance behaviour is where Compliance Theatre lives.
Q7 Does your leadership team actively sponsor AI governance — attending reviews, making resourcing decisions, holding people accountable — or has it been delegated entirely to a compliance or technical function?
Q8 How many AI-related incidents or near-misses has your organisation formally logged and reviewed in the last 12 months?
Analysing governance exposure
Mapping responses to risk dimensions...
— Exposure band —
Your governance exposure diagnostic is ready.
Your responses have been scored across four risk dimensions: Shadow AI inventory, oversight coverage, human capability, and governance culture. Enter your work email to see the full breakdown — including the specific gaps your answers flagged.
Please enter a valid work email address.
Work email only. Your results will also be emailed to you. No spam — Vivadence is a professional service, not a marketing platform.
—
—
—/24
—
Risk dimension breakdown
Specific findings from your responses
Next step
The AARI Assessment tells you what's actually there.
This diagnostic is a directional signal. The AARI Index — completed with a minimum of 5 respondents across your designated oversight team — produces a validated, audit-grade measurement of human governance capability across 7 dimensions. It is the only instrument that produces evidence acceptable under Article 14 of the EU AI Act.
This is a directional diagnostic, not a formal AARI assessment. Scores are based on self-reported single-respondent answers and do not constitute audit-grade evidence. The AARI Integrated Method requires a minimum of 5 respondents per AI system assessed and produces output validated against documented compliance evidence. Vivadence Limited · Patent Pending · Confidential
