Trust Center
AI governance
Governance is only credible when each principle shows up in the product. Here is how Omniday implements six of them.
Transparency
Every AI-facing surface discloses that the user is interacting with an AI. The chat widget carries the disclosure in its header and greeting, in line with EU AI Act Article 50.
Human oversight
Your organisation defines where the AI may act on its own and where a person must approve. AI-drafted email replies require human approval before sending by default.
Traceability
Answers can reference the knowledge sources they were grounded in, and enquiries become structured cases with a record of changes, approvals and actions.
Data minimisation
Chat messages are scrubbed of personal identifiers before export to observability tooling, and retention is defined by your organisation rather than defaulting to 'keep everything'.
Testing
Assistant behaviour is covered by automated scenario tests — including boundary cases, escalation rules and tenant isolation — that run before changes ship.
Accountability
Omniday acts as a data processor under a GDPR Article 28 agreement, with a named sub-processor register and 30 days' notice of changes.
On the EU AI Act
Omniday is designed to support the transparency, logging, human-oversight and documentation practices the EU AI Act expects of AI-assisted customer service. We do not describe Omniday as “AI Act compliant” — compliance depends on your concrete deployment, and your organisation remains responsible for assessing it. We provide the controls and the documentation to make that assessment straightforward.
Questions about a specific deployment? Contact hello@omniday.ai or review the data-location matrix.