Guide / AIGIS resources
EU AI Act obligations phase in across 2026, with a major set of rules effective 2026-08-02 and further rollout into 2027. AIGIS does not certify compliance; it produces permission-provenance evidence that supports AI control and oversight conversations. Salesforce is the production proof path.
Executive read
Treat EU AI Act readiness as an evidence workstream, not a certification claim.
Capture what the model was allowed to see before it answered, and what was excluded.
Keep an auditor-readable record of permission decisions per AI interaction.
Start on a Salesforce workflow with permission-sensitive data; scope ServiceNow and SAP as disclosed co-development.
Analysis
EU AI Act rules apply progressively, with a significant set of obligations effective 2026-08-02 and further rollout foreseen into 2027. Teams with EU exposure need repeatable evidence now, not a policy binder later.
AIGIS produces evidence useful for those conversations. It does not guarantee legal compliance, and no AIGIS material should claim EU AI Act certification.
For each governed AI interaction, AIGIS records the permission checks performed before model exposure and the fields or records excluded from the prompt.
That permission-provenance record is what lets a security or compliance reviewer see why the AI was allowed to know what it knew.
Pick one Salesforce workflow with role-dependent visibility. Show object, field, and live record access checks before prompt construction, then the evidence trail after the answer.
Book a scoped governance review at `/demo` to map this to a real workflow with EU exposure.
Keep reading
Security
A concise CISO-oriented brief for evaluating AI governance evidence, permission enforcement, and connector asymmetry.
Security
A regulated-industry lens on permission provenance, field stripping, and Salesforce-first proof workflows.
Audit
The audit elements needed to explain what an AI answer was allowed to know.