From playbooks to decisions
Security orchestration, automation, and response (SOAR) tools run playbooks that people write in advance. They execute reliably, but someone still has to decide which playbook fits an alert, and every new case means another playbook to build and maintain.
AI SOAR adds a reasoning step before execution. AI works out whether an alert needs a response and which action fits. A deterministic engine then carries out that action the same way every time, within the limits you set.
How AgentSOAR fits into AgentSOC
AgentSOAR is a module of AgentSOC, not a separate product. Triage and correlation decide what needs a response, the response agent builds a remediation plan, and AgentSOAR executes it against your connected providers. Each execution moves through clear states (pending, running, succeeded, failed, or reverted), and the incident report picks up the result.
Cloud credentials are encrypted with AES-256-GCM and mapped to the domains they cover, so each action runs with the right account. AgentSOAR also syncs an inventory of hosts, users, and IPs from your cloud providers, so actions target real assets.
Bring the automation you already run
If your team already has workflows in Shuffle, connect your own instance and trigger them from Jutsu investigations. Connectors for Tines, Splunk SOAR, Cortex XSOAR, n8n, and Torq are on the integrations roadmap.