Skip to content

AI-Powered Playbooks

Bedrock generates contextual, step-by-step remediation guidance for the specific threat (advanced tier and above). Every identifier comes from the actual threat record — the model fills placeholders, never invents resource names or ARNs, and any output that does is rejected.

This is the deterministic framing applied to prose: the AI explains, it never decides. The threat score, count, and severity are computed by rules and math; Bedrock writes only the narrative of what to do about it, with numbers templated in.

On-demand generation

A playbook is created the first time you open a threat and click Generate Playbook, then cached and reused. This ties Bedrock usage to the threats you actually work rather than to detection volume — a burst of many distinct threats does not run up model cost or back up a generation queue. The first view generates in seconds; every view after is instant.

Privacy: identifiers redacted before AI, encrypted at rest

Identifier redaction before the AI is always on, on every tier, with no setting to turn it off:

  • Real identifiers (IAM principals and ARNs, IP addresses, account and resource IDs) are replaced with type-preserving placeholders before anything is sent to Bedrock — the model sees <PUBLIC_IP> or <ROOT_PRINCIPAL>, not the literal value — and the real values are restored only when the result is shown to your analyst.
  • The cached playbook is stored with its identifiers redacted, and the lookup map is encrypted with a dedicated KMS key in your account, so real identifiers are never written to the AI cache in the clear.
  • Before a generated playbook is stored, it is verified to contain no un-redacted real identifier. If anything slipped through, the model output is discarded and a safe deterministic template is used instead, with an operational alarm raised. The deterministic fallback is itself also redacted and encrypted at rest.
  • Each playbook shows a privacy provenance badge — redacted before AI & encrypted at rest, or a plain-template fallback — and the badge is a verified claim, re-checked against the stored playbook on open, rather than a stored marker.

This is data minimization layered on top of an already in-account, no-retention Bedrock call.