AI policy gap register
- Gaps: 0 systems, 0 decisions a month.
- Compliant: 0 systems.
Global Compliance · Expanded Global Compliance
AI Governance Assurance
Keep an inventory of the customer's AI systems and agents: model and provider dependencies, risk assessments, use cases, approvals, policy obligations, human oversight,…
Demonstration: Meridian Freight Group (fictional) · Freight and logistics. Fictional data and a scripted answer.
The situation
- Organisation
- Meridian Freight Group (fictional)
- Who is asking
- Head of AI governance
- Industry
- Freight and logistics
- What happened
- Teams have started using AI tools, and the governance committee wants to know which systems meet the company's own AI policy.
The question asked
Which AI systems in use fall short of our AI policy, and how many decisions do they influence?
What it reads, and who can see the result
- AI system register registerSystems with owner, purpose and assessment status.
- AI policy documentThe company's requirements.
- Reads
- The AI register entries and the policy.
- Does not read
- Model internals or the data processed.
- Permission
- The AI governance office's register access.
- Who sees it
- The head of AI governance and the AI committee.
Try it
Review each flagged record: confirm it or dismiss it. The totals and the answer below follow your decisions.
| System | Used for | Owner | Decisions a month | Finding | Decision |
|---|---|---|---|---|---|
| Route optimiser | Planning delivery routes | Operations | 18,000 | Matched | None needed |
| CV screening assistant | Shortlisting applicants | None recorded | 1,400 | No owner, and it influences hiring decisions without human review | |
| Invoice reader | Extracting invoice fields | Finance | 9,200 | Matched | None needed |
| Customer chat assistant | Answering customer questions | Customer service | 22,000 | No impact assessment; the policy requires one for customer-facing use | |
| Fraud scoring model | Scoring payment risk | Finance | 6,400 | Human review required by policy is not evidenced | |
| Meeting summariser | Summarising internal meetings | IT | 3,000 | Matched | None needed |
- Flagged
- 3 of 6 · 29,800
- Policy gap
- 0 · 0
- Compliant
- 0 · 0
- Undecided
- 3 · 29,800
The answer
Simulated answer. Written for this fictional case, not produced by a live model. Only the figures in it are computed, from what you decided above.
3 of 6 AI systems fall short of the policy, influencing 29,800 decisions a month.
Three gaps appear: a system with no owner that influences hiring, a customer-facing assistant with no impact assessment, and a scoring model whose human review cannot be evidenced [S1][S2]. The first is the most serious: a hiring decision with no owner and no review.
You have recorded 0 as policy gaps (0 decisions a month) and 0 as compliant (0); 3 are undecided.
Relies on: S1 · AI system registerS2 · AI policy v1
Review and approval
The head of AI governance decides. Approving sends remediation requests to the system owners for the gaps recorded. It does not switch any system off.
Not ready to approve. 3 flagged systems are still undecided. Confirm or dismiss each one first.
In the product this is a different person with the right permission. Here you play that role to see the rule.
What it produces
Limits of this example
- Checks the register against the policy; it does not test how a model behaves.
- A system missing from the register is invisible to this review.
- This is a scripted demonstration on fictional data. No model was called and no real organisation, person or result is shown.
More in Global Compliance
- The compliance command centreObligations, deadlines and changes are scattered across inboxes and spreadsheets, so something is eventually missed.
- Corporate and financial regulationFilings, tax obligations, licences and entity records are tracked by different people, often on different dates.
- People, sector and international dutiesEmployment, privacy, security, environmental and cross-border duties sit with different teams and are rarely seen together.
- Assurance and evidenceWhen an auditor or an inspector arrives, evidence has to be found in a hurry, and gaps appear for the first time.
- Regulatory Applicability IntelligenceWork out which obligations may apply to this business from its legal entities, activities, industry, locations, workforce and other verified facts, and show each as appl…
- Regulatory Change Impact IntelligenceWatch approved authoritative sources, detect changes that matter, identify the entities, processes, systems and controls they affect, and start a governed review and rem…
- Continuous Control MonitoringEvaluate whether each control is operating: evidence present and current, expirations, exceptions and failures, so assurance is ongoing instead of a one-time checklist.
- Contract Compliance IntelligenceRead customer-authorised contracts for obligations, service commitments, deadlines, renewals, restrictions, responsibilities and compliance risks, with citations to the…
The Global Compliance story · This capability in the full register