AI & Agents

Multi-Agent Harness for IT Operations

Specialised agents behind one super-agent, acting on enterprise systems through an MCP gateway, inside guardrails, routing, audit and a hand-over to people.

A reference multi-agent harness for IT operations. Requests from the ITSM desk and chat pass one gateway with single sign-on and guardrails; a super-agent routes each job to the specialised agent that owns it. Agents reach models only through an LLM router, reach enterprise systems only through an MCP gateway with a policy check on every call, and draw on a vector store and a knowledge graph. When a procedure runs out, an engineer takes over.

ticketsrequeststoken checkmodel callsknowledge · memorypolicy checksignalsopens a tickethand-overIdentity, models, runtime and policyWork coming inGateway pipelineAgent workflowsSpecialised agentsToolsEnterprise systemsKnowledge, memory and eventsAuditPeopleKeycloakSSO · SAML / OIDCAD / LDAPcustomer directoryLiteLLMmodel routing · budgetsLLMsClaude · OpenAIKubernetescontainer orchestrationAWS Secrets ManagerOPApolicy as codeITSM deskticketsChat and portalusers, IT operatorsMonitoring signalsapps, databases, servicesKongAPI gatewayGuardrails AIinput and output checksSuper-agentroutes by intentIntake agentclassify, routeKnowledge agentanswers with sourcesData agentquery → run → checkMonitoring agentdetects anomaliesRunbook agentconfirm, guide the fixChange agentplan → deploy → testMCP gatewayContextForgeTickets and changesITSM recordsDatabasesServers and legacyappsAPI · MCP · DBKafkaevent modelQdrantvectors · BM25 + RRFNeo4jknowledge graphPostgreSQLoperational dataMongoDBconversations, contextRedisshort-term memoryObservabilitylogs · traces · alertsEngineerhand-over · approvals

Point at a component to light its links; select it for what it does here and the practice that makes it right on AWS.

When to use it

  • An IT operations desk with a high volume of routine requests, data queries and change work
  • Monitoring signals that should open and work tickets before a person notices
  • Agents that must act on real systems with identity, policy and an audit trail

How NGT delivers it

Built from code, reviewed at a gate by an NGT architect, and handed over with runbooks your team can run.

AI & Agents