FPT SoftwareFHNCertificationsReference architectures
Talk to NGT
AIOps · 03 of 5

Agentic IT operations

Specialised agents behind a super-agent; monitoring hands anomalies straight to its agents.

Agentic IT Operations
01

Specialised agents behind a super-agent

A reference pattern for multi-agent IT operations. Walk it from request to fix.

Agentic IT Operations
routesanomalyno ticketguided fixtool callsevery call checkedWhere work comes inMulti-agent platformSpecialised agentsSecond path · no ticket raisedSystems the agents work onControls around every agentChatplain-language asksITSM ticketsservice-desk intakeChange requestsconfig · releasesMonitoringapps · DBs · servicesSuper-agentroutes · keeps contextIntake agentrouting · priorityKnowledge agentanswers with sourcesData agentquestion → query → checkChange agentplan → deploy → test → reportMonitoring agentwatches · detectsRunbook agentconfirms · guides fixMCP gatewayregistered tools onlyEngineerapproves · takes overITSMtickets updatedMonitoringmetrics · logs · alertsDatabasesgoverned queriesInfrastructureconfig · capacityCI/CDdeploy · testIdentity & gatewayKeycloak · KongModel routingLiteLLM · Claude · OpenAIGuardrails & policyGuardrails AI · OPAKnowledgehybrid RAG · BM25 + RRFAudit trailevery action loggedAgent monitoringthe agents watched too

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

Point at a component to see what it does; select it for the best practice.Open the architecture
02

Two everyday flows, before and after

Before: manual tickets and L1/L2 hand-offs. After: agents carry routine work; people step in for judgement.

Before With agents
Logging a requestThe user logs a ticket by hand.The user asks in chat; the agent captures context, at any hour.
RoutingL1 receives the ticket and forwards it to L2.The intake agent routes it; FAQs and user errors are solved on the spot.
Data requestsL2 writes and runs the query by hand.The data agent writes, runs and checks the query.
MonitoringAdministrators check several monitoring tools.The monitoring agent runs the checks and summarises.
Abnormal behaviourL1 logs a ticket once someone notices.The monitoring agent hands it to the runbook agent; nobody raises a ticket.
FixingL1 searches documents for the procedure.The runbook agent confirms the fault and guides the engineer.

How the agents are built and governed

AI & Agents: the agent harnessMLOps: releasing models with sign-off
SourcesNGT AI-First DevOps Transformation service offering (Oct 2026)NGT Services Offering v2.0 (Mar 2025)NGT Cloud Operation service offering v0.2 (Jan 2025)NGT Cloud, DevOps & SRE capability deck (Sep 2026)FHN AI Transformation Service (Aug 2025)FHN AI Service Introduction v0.8.5 (Nov 2025) and v0.9.3 (Mar 2026)NGT engineer profile sample, 38 CVs (2024–2026)