← ShiftPilot AI

Security

This page is maintained by ShiftPilot AI to answer common security questions from dealerships evaluating our platform. It describes controls that are enabled today — not a certification.

Tenant isolation

Every dealership's data is isolated at the database level using Postgres Row-Level Security. Users can only read and write rows scoped to a dealership they belong to.

Role-based access

Access to sensitive workflows (finance, deal desk, customer PII, integrations, exports) is gated by role: dealer principal, general manager, sales manager, sales rep, finance manager, service manager, service advisor, accounting, and read-only.

Server-side workflows

AI model calls, lender submissions, document generation involving sensitive data, exports, and admin overrides run on the server. Frontend code never touches provider secrets.

Audit logging

Sensitive actions — role changes, exports, integration connect/disconnect, support-mode access, deal and finance updates — are recorded to an append-only audit log that regular users cannot modify.

AI approval workflow

ShiftPilot AI agents are assistive, not autonomous. High-risk actions (sending customer messages, submitting finance data, pricing changes, exports) are stored as pending and require a manager approval before execution.

Export controls

Exports of customer or finance data require the correct role, are rate-limited, are redacted by default, and are recorded in the audit log. Download URLs expire.

Data at rest and in transit

All connections use TLS. Application data is stored on managed Postgres with encryption at rest provided by the cloud host.

What is not claimed

ShiftPilot AI does not claim SOC 2, ISO 27001, PCI, HIPAA, or FTC certification. Production deployments processing live consumer finance data should complete legal review, a security review, and a third-party penetration test before go-live.

Contact

Security disclosures: security@shiftpilotai.io