Skip to content

Where to Begin

Where you begin depends on what you are doing now. Five entry points, one triage.

Pick your situation

  • I am shipping an LLM feature in days, not weeks

    Seven controls, one checklist, one decision tree. Deliberately short, deliberately opinionated. Once these are in place, you have a Minimum Viable AIRS deployment: enough runtime safety to go live, enough observability to learn, enough structure to decide where to invest next.

    AIRSLite

  • I am deploying low-risk AI and want to move fast

    Pre-approved controls for low-risk deployments. If your system qualifies (internal, read-only, no regulated data, human in the loop), you deploy without a formal security assessment. If it does not, the Fast Lane tells you exactly which path to take instead.

    Fast Lane

  • I need working code in thirty minutes

    End-to-end: classify risk, deploy guardrails, configure a Judge model, set up human oversight. Opinionated setup, runnable examples, zero multi-quarter runway required.

    Quick Start

  • I want to understand the framework before I build

    Curated reading paths through the full framework, organised by goal. The Golden Thread takes you from why runtime security? through each control layer to how the system self-corrects in about two hours. Other paths cover threat landscape, multi-agent security, and practical artefacts.

    Start Here (Reading Paths)

  • I have a specific question about scope, cost, or standards

    Honest answers to the questions practitioners actually ask: does this apply to vendor AI, what does a Judge cost, how does it map to ISO 42001, what is the minimum team size to operate this.

    FAQ

If you are still not sure

Walk through these in order. Each one answers a question the previous one raises, and you can stop at any point once you have what you need.

  1. AIRSLite gives you the short list. If the seven controls are enough for your situation, you are done.
  2. Quick Start turns the short list into working code.
  3. Reading Paths gives you the reasoning behind every control, so you can adapt them to your context rather than copying them blindly.
  4. For Your Role tells you what any of this means for your specific responsibilities: security leader, product owner, AI engineer, risk manager, and six others.

Prefer a role-based path?

If you would rather triage by who you are than by what you are doing, jump straight to For Your Role. Each page tells you what matters for your role, why, and where to start reading.

References