SAML SSO, two-factor authentication, six-level role-based access control, IP allowlisting, and a queryable audit trail, with automatic PII detection and encryption in transit and at rest. Security isn't an add-on — it's in the foundation of every plan.
Most CX platforms route every customer conversation through shared AI infrastructure with no contractual guarantees about data retention, model training use, or geographic residency. You sign up for a support tool and discover your customers' data is in 12 subprocessors you've never heard of. If that's a compliance risk for your industry, you need a platform that gives you control.
Without a comprehensive audit log, you can't answer: who viewed this conversation? Who changed this customer record? What did the AI say to this customer at 3 a.m.? When a regulator asks, or when an incident happens, most platforms leave you with no evidence. Compliance isn't possible without observability.
When a customer exercises their right to erasure, your support team files a ticket, a developer writes a one-off SQL query, someone checks if vector embeddings were purged, and nobody is sure the job is complete. Manual deletion pipelines fail silently, expose you to regulatory risk, and can't be audited after the fact.
Without role-based access control, a new hire has the same data visibility as your Head of Support. HIPAA, PCI-DSS, enterprise client contracts, and basic operational security all require need-to-know access controls. Flat permission models aren't just inconvenient — they're a liability.
Logins, admin and role changes, conversation assignment and closure, knowledge-base edits, and integration connections are written to an audit log with timestamp, user, and details. Org admins view their organization's trail from Settings; platform admins can view activity across every organization.
AItocha scans inbound messages for credit card numbers, Social Security numbers, phone numbers, and other PII patterns in real time, and can redact them before storage. Wired into live chat, voice, and inbound webhooks.
Contact records can be erased on request directly from the platform — no developer ticket required for the standard case.
Owner, Admin, Manager, Supervisor, Agent, and Viewer — each with a defined permission set covering conversations, analytics, team management, queues, content, billing, and leads, enforced on every request rather than hidden in the UI.
Every API key carries its own read/write/admin scopes, so an integration gets exactly the access it needs. Revoke a single key without rotating every other integration.
Connect Okta, Azure AD, or any standards-compliant SAML 2.0 identity provider. Owners keep password login as a fallback even when SSO is enforced for everyone else, so a misconfigured IdP can never lock your team out.
TOTP-based 2FA with any authenticator app (Google Authenticator, 1Password, Authy), plus single-use backup codes if a device is lost. Configured per user from Settings.
Restrict sign-in to specific IP addresses or CIDR ranges. Checked at login, not on every request, so a bad entry only blocks new sign-ins — never an already-active session.
Point your AI bot at a private or self-hosted model — vLLM, Ollama, or any OpenAI-compatible endpoint — so inference runs on your own infrastructure instead of ours.
| Auto PII detection & redaction | Yes |
|
| GDPR erasure from the platform (no dev needed) | Yes |
|
| Per-key scoped API access | Yes |
|
| SAML SSO without a per-seat add-on fee | Yes |
|
Not yet — we haven't completed a SOC 2 audit. We can walk your security team through the controls we do have (RBAC, SSO, 2FA, audit logging, PII detection, encryption in transit and at rest) on a call.
Contact records can be deleted directly from the platform. For anything outside the standard case, contact support.
Logins and logouts, admin and role changes, conversation assignment and closure, knowledge-base edits, integration connections, and a few admin actions are currently logged, each with timestamp, user, and detail. We're expanding coverage over time — ask us about a specific event type if it matters for your compliance review.
Via role-based access control. Six roles are available — Owner, Admin, Manager, Supervisor, Agent, Viewer — each with different permissions across conversations, analytics, settings, and billing. Admins can further scope an agent to specific channels, queues, sites, or teams.
From Settings > Security, enter your IdP's entity ID, SSO URL, and signing certificate, then hand your IdP admin the ACS URL and metadata URL we generate. You can optionally require SSO for everyone except owners, who always keep password login as a recovery path.
Contact support before or shortly after cancellation to arrange a data export or deletion timeline for your account.
No — your conversation data is not used to train third-party foundation models.
SSO, 2FA, role-based access control, audit logging, PII detection, and encryption come standard on every paid plan — no per-seat security tax. Tell us what your security review needs and we'll tell you plainly what we do and don't have today.
Start Free Trial Blog