AI Use Policy

AEGIS OS is a governed operating system. Intelligence assists execution inside defined boundaries; it does not replace accountable human decisions. Every AI role in a deployment is defined with an owner, an approval boundary and a record of what it did.

• Anything that commits money, signs, or binds the business requires human approval.
• Anything sent externally under a client’s name is reviewed before it leaves.
• Anything that changes a system of record is logged and attributable.
• Legal, clinical, credit and investment decisions are not automated.

Client deployments use commercial model providers selected per engagement. OpenAI is the primary provider, with Anthropic and Google Vertex available as governed alternatives where an engagement calls for it. The providers used in a specific deployment are documented in that engagement’s architecture.

Client data used in a deployment is governed by that engagement’s agreement, including which data may reach a model provider and which must not. We do not use client data to train public models.

On this website, information you submit through the contact form or the pre-assessment is stored for follow-up and is not sent to a model provider. Website copy on GetAegisOS.com is written and reviewed by people before publication; resource articles are published only with a named author and review.

We do not deploy intelligence that impersonates a person, hides that it is automated where that matters, operates without a named owner, or acts outside an approval boundary agreed with the client.

Email mail@getaegisos.com.