Not based on a specific customer deployment. The company, figures, decisions, and outcomes below are a constructed example that shows how Intended’s mechanism applies to this kind of problem. They are not measured results from a named customer. Where we publish a real, attributed customer outcome, we will say so explicitly.
Mapping AI Governance to the EU AI Act in the Public Sector
Consider a public-sector agency whose AI agents handle document processing, case management, benefits eligibility determination, and citizen communication — operating under transparency and data-residency obligations such as the EU AI Act.
01 · the challenge
The kind of problem this addresses.
02 · how it works
See the difference.
Citizen submits benefits application
Application #FA-2026-48291
AI agent evaluates eligibility
No transparency into decision logic
Decision rendered
Denied -- no explanation provided to citizen
No audit trail
Cannot reconstruct why AI denied application
03 · the solution
What they deployed.
- — Policy rules for citizen-service and case-management actions, written by the agency
- — Policy requires a caseworker's review for every adverse decision, the human oversight the EU AI Act expects for high-risk systems
- — Policy denies agent actions that would send citizen data to a system outside the approved jurisdiction, by matching the destination in the request
- — Each decision record carries the policy and the reason, which caseworkers use to write the citizen's explanation
- — Tamper-evident, hash-chained audit records, plus a signed receipt per decision that verifies against the agency's public keys
- — Agencies that cannot use the managed cloud can ask about single-tenant and air-gapped deployment, available on request
04 · implementation
From zero to governed.
Phase 1
Assess
Map AI agent decision points against EU AI Act requirements. Identify the high-risk AI use cases that require governance.
Phase 2
Configure
Write the human-review, jurisdiction, and citizen-privacy rules. Choose the managed cloud or an on-request self-hosted deployment.
Phase 3
Integrate
Add the SDK call ahead of agent actions in the case-management, benefits, and document-processing systems. Train caseworkers on the approval queue.
Phase 4
Evidence
Run the compliance evidence export. It scores each EU AI Act control it can support as satisfied, partial, manual, or not covered, as input to the agency's own compliance review.
05 · illustrative outcomes
What this is designed to deliver.
Modeled figures for this scenario — what the workflow above is built to achieve, not measured results from a named customer.
0%
Adverse decisions human-reviewed
By design, per EU AI Act mandate
0%
Decisions explainable
Citizen-facing explanation from the authority trace
0%
Out-of-jurisdiction exports denied
By policy on the destination system
0
Control statuses in the export
Satisfied, partial, manual, not covered
06 · decision replay
Example decisions, full trace.
Sample decision records that show the shape of the evidence Intended produces. Illustrative, not drawn from a live customer’s logs.
gov.citizen.benefits-determinationRISK: 76/100ESCALATEAI evaluates housing assistance eligibility for Application #FA-2026-48291
Resolved by: Senior Caseworker (reviewed and confirmed the denial)
gov.document.classificationRISK: 22/100ALLOWAI classifies submitted tax documents for verification
Resolved by: Policy: document classification auto-approved (non-adverse)
gov.citizen.communicationRISK: 34/100ALLOWAI generates response letter to citizen inquiry about application status
Resolved by: Policy: informational communications auto-approved with template check
gov.citizen.data-exportRISK: 100/100DENYAI agent attempts to export citizen PII to analytics system outside jurisdiction
Resolved by: Policy: data residency violation -- target system outside sovereign boundary
the takeaway
The EU AI Act requires that citizens can understand and challenge AI decisions that affect them. Mapping authority traces to that requirement provides explainability and human-oversight evidence without rebuilding the underlying AI infrastructure.
Why this pattern matters — not a customer quote.
Start achieving AI compliance
Free to start. See every AI decision from day one.