What’s built, plainly.
- RLS on every table
- Security-audited (critical/high fixed)
- Data residency in Canada (ca-central-1)
- PIPEDA/PIPA alignment
- Audit trail on every action
- Per-employee sensitivity shielding
- Consent auto-expiry for candidates
- Plain-English export, no lock-in
French-language support is on our roadmap, designed for Bill 96 compliance.
AI trust
What can Ask Peopl do without a human?
All actions move through your existing approval process. The AI drafts the work, and nothing proceeds until a human approves it. The workflow is intentionally designed to keep people at the center.
Does customer data train AI models?
No. Customer data does not train AI models.
Which AI providers are used and where is data processed?
Your records are stored in Canada (ca-central-1). AI features send the relevant context to commercial model providers, which process it in the United States under API agreements that prohibit training on your data. Prefer to keep AI under your own control? Bring your own AI: connect your company's provider over the Model Context Protocol, so AI requests run under your agreement and your keys.
How does the audit trail record AI-initiated actions?
Every AI-staged change lands as a request with the actor, the timestamps, and the full approval chain, the same trail as a human-submitted change.
- 09:14:02Ask PeoplAIComp change staged for approval
- 09:14:31Dana ReyesComp change approved
- 09:15:08Ask PeoplAIflaggedBiased phrasing flagged in feedback
- 09:16:00SystemLetter filed to the employee record
Human and AI actions share one trail, with the actor, the timestamps, and the approval chain.
Certifications
SOC 2 certification is on our roadmap. In the meantime, we’re happy to walk through our security practices directly. Ask us anything.
Accessibility
Peopl.ca targets WCAG 2.1 AA. Found an issue? Tell us.