Security is how we build, not a page we wrote.
The same rules apply to our products, the tools we run and the software we build for clients: isolate, encrypt, log, back up, rehearse.
01 — Controls
Identity first
Single sign-on, multi-factor authentication and role-based access on everything we run. Conditional access where the tool supports it.
Tenant isolation
Row-level security in our products. Separate environments for managed tools.
Encryption
TLS everywhere, encrypted storage, encrypted off-site backups. End-to-end encryption in Hamraah.
Monitoring and response
Around-the-clock monitoring on enterprise agreements, defined incident playbooks, evidence preserved.
Audit trails
Immutable postings in the ERP, hash-chained events in Cairn, audit logs in every tool.
Secure delivery
Branch protection, reviewed changes, secrets in vaults, dependency audits on every release.
02 — AI
How we handle AI
AI features in our products and the assistants we deploy for clients run with permissions, not around them. An agent gets the context it is authorised to read, proposes rather than acts by default, and every call is logged.
Client data is not used to train models. Where a private inference stack is needed, we build one inside your boundary.
Need documentation for procurement?
The Trust Center has policies, subprocessors and disclosure details. We'll answer a questionnaire too.