Tamnoon’s Tami platform uses specialized agents and safety checks for cloud remediation; the company reports up to 90% lower critical exposure in 90 days.
AWS described Tamnoon’s Tami cloud-security platform on October 14, 2026, detailing how it uses Amazon Bedrock to investigate security findings and prepare or carry out remediation. The platform ingests alerts from cloud-native tools and cloud security products, adds asset and exposure context, then routes work through specialized agents. Tamnoon says Tami has analyzed more than 14 million cloud alerts across its customer base.
Four agents handle separate stages. An investigation agent gathers context and produces read-only investigation steps. A safety-scoring agent classifies proposed remediation as SAFE, RISKY or AWAITING DATA. A remediation agent creates a plan only for findings rated safe, with backup, verification and rollback steps. A code-generation agent turns approved plan steps into executable changes; a separate reviewer agent must approve each generated artifact.
The system is designed to keep planning separate from execution. The AI does not hold cloud credentials, according to AWS’s account of the architecture. Changes are executed using short-lived, tenant-scoped credentials, with approval workflows and audit trails. Guardrails screen requests and responses, while human review is available when the system cannot reach a reliable result. Tami connects with detection products including Wiz, CrowdStrike and Palo Alto Cortex, as well as AWS Security Hub and Amazon GuardDuty.
AWS attributes the reported outcomes to Tamnoon’s research. For a typical customer, the company reports up to a 90% reduction in critical cloud exposure within 90 days, up to 100 times greater remediation efficiency through root-cause clustering, and zero production issues across millions of workloads. Tamnoon also reports more than 13,000 engineer hours saved per week across its customer base. These are company-reported results, not independently established benchmarks.
The implementation runs in Tamnoon’s AWS environment and uses Bedrock Cross-Region inference within the US geographic boundary, AWS says. Bedrock Guardrails, private network placement, least-privilege permissions and tenant separation form additional controls. Tamnoon positions Tami as a remediation layer that works alongside existing detection tools rather than replacing them.