// SECURITY & DATA SOVEREIGNTY

Security & data sovereignty

Self-hosting means the AI model runs on hardware you control — a workstation, a rack in a local datacentre, or your private cloud — instead of a metered API from OpenAI or Anthropic. Your prompts and documents never leave your infrastructure, which makes privacy reviews dramatically simpler under the AU and NZ Privacy Acts.

Default security posture

No internet call required at runtime

Default posture: models and data live on your infrastructure. Air-gapped deployments are fully supported.

SSO + per-user access controls

Access is per-user via your existing SSO, and every agent action is written to an audit log.

Human-in-the-loop by default

Agents start supervised and earn autonomy gradually, with human-in-the-loop approval gates you control.

Privacy Act alignment by design

Deployments align to the Australian Privacy Act and New Zealand Privacy Act 2020 by design, not by questionnaire.

Data residency

Models run on customer-controlled hardware in Australia or New Zealand — a workstation in your office, a rack in a Sydney, Melbourne or Auckland datacentre, or your own cloud tenancy. Your VPC, your tenancy, your keys.

There is no offshore LLM routing by default: prompts, documents and model weights stay on your infrastructure, and no customer data crosses a public API on the data path.

BOOK A WORKLOAD AUDIT

Book a 45-minute workload audit. We'll map where agents genuinely pay off, what the hardware costs, and if it's not the right move, we'll say so.

➤ hello@agentos98.ai