Built for controlled AI operations.
Designed with tenant isolation, approval trails, and role-based access at the core. The platform is structured around the principle that operators stay in control of what AI executes on their behalf.
Four operating principles.
Every architectural choice traces back to these four. They're the lens we use when we add a feature, ship an integration, or decide a default.
Risky actions never auto-execute. Autonomy widens only when you explicitly say so, per class, per workspace. The agent prepares; the human decides.
Tenant boundaries are enforced at every layer — data, compute, identity, and AI context. Agency and client workspaces are separated by default, not as an opt-in.
Every agent action ships with its reasoning. Approvers see the why, not just the what. The audit trail is immutable and exportable for compliance and incident review.
Policies set at the agency tier cascade into every client workspace. Overrides are explicit, attributable, and logged — you always know who allowed what.
Tenant isolation at every tier.
A simplified view of how data, identity, and AI context separate across the agency / client / contact tiers.
AGENCY TIERPARENT
CLIENT TIERCHILD
CONTACT TIERDATA
DATA NEVER CROSSES TIERS WITHOUT AN EXPLICIT, LOGGED POLICY OVERRIDE.
Controls you can configure today.
Tenant isolation
Per-workspace data boundaries. Agency and client separation enforced at every layer.
Role-based access
Owner · approver · contributor · view-only. Per-resource overrides for fine-grained control.
Audit trails
Every agent action logged with reasoning. Immutable. Exportable as CSV or JSON for compliance.
Manual-first autonomy
Risky actions never auto-execute. Autonomy widens only when you say so — per class, per workspace.
Data privacy controls
Field-level redaction, configurable data retention, and per-tenant encryption keys on the roadmap.
Activity log & reasoning
Each decision the agent makes shows its rationale. Approvers see the why, not just the what.
Cascading governance
Approve policies once at the agency tier; cascade into every client workspace with override-on-purpose.
White-label governance
Custom branding, domain, and policy boundaries respected end-to-end — including in client-facing surfaces.
On certifications & standards.
MarketinQ is in early access. We aren't claiming certifications we haven't earned — but the platform is being designed from day one against the controls those certifications expect.
What standards is MarketinQ designed against?
Architecturally, we target the controls in SOC 2 Type II, ISO 27001, and GDPR / CCPA from the start — tenant isolation, role-based access, encryption in transit and at rest, audit logging, data subject rights handling. Formal audits will follow once we've completed the early access cohort.
Where is data stored?
Primary infrastructure runs in tier-1 cloud regions with regional residency options on the roadmap for white-label tenants. Backups are encrypted, regionally redundant, and retention is configurable per workspace.
How is AI access to data controlled?
The agent operates within the workspace it's invoked from. It cannot read data outside that workspace, cannot persist context across tenants, and every action it takes is logged with the inputs it saw and the reasoning it produced.
Can I export an audit trail?
Yes — CSV and JSON exports are available at any time. The trail is immutable and includes the agent's reasoning, the approver's identity, and the resulting state change.
What about data deletion and the right to be forgotten?
Contact-level deletion is supported in-app, with cascading removal across associated message history and segments. Workspace-level deletion is available on request and follows a documented process to preserve required compliance records.
Talk to us about your controls.
We're happy to walk through our architecture, share our security questionnaire responses, and discuss your specific compliance posture.