L1/L2 support
Resolve more at L1. Keep engineers engineering.
Known issues are matched to your runbooks and closed at L1. The rest go to the right team's agent first responder, which finds the root cause before L2 picks it up.
L1 resolution
Known issues close at L1.
NOFire AI recognises the failures your team has seen before and matches each one to the right runbook. L1 follows the steps and closes the ticket. Up to 60% of L1 tickets are resolved without escalating.
- Your runbooks, put to work. NOFire AI reads the runbooks you already have. Nothing new to write.
- The common failures, recognised. Repeat issues like a failing third-party API, a full disk or a stuck job are spotted on arrival, before anyone escalates.
- Every fix is recorded. Each resolution is logged with the runbook step that fixed it, so the next one goes faster.
L2 escalations
The first responder isn't a person anymore.
When L1 can't close a ticket, the owning team's agent takes it first and digs in. Support specialists step in only when needed.
- One agent per team. Set up with the team's services and runbooks.
- Routed to the owner. The ticket reaches the team that owns the failing service.
- Looks for itself. Deploys, metrics and logs, checked before anyone is paged.
L3 handoff
Engineers start from the answer.
Escalations arrive with the cause, what was ruled out and what was tried. Support work takes around 20% of every sprint. Starting from the answer gives that time back.
Your stack
Works with the tools your teams use.
Runbooks
Confluence
Notion
Google Docs
GitHub
Ticketing
Jira
ServiceNow
Zendesk
Chat
Slack
Microsoft Teams
Engineers' tools
Claude Code
Cursor
NOFire CLI
MCP
Start where you want
Start with L1. Add the rest when you're ready.
Begin with low-priority tickets, then add L2 and more teams.
Frequently asked questions.
What if our runbooks are out of date, or we don't have many?
Start with what you have. NOFire AI checks each runbook against the live map of production and rules out the ones that no longer match what is running. When an incident is resolved, it drafts a runbook from what fixed it, and a person approves it.
What access does the agent need?
Read-only roles in your cloud, cluster and monitoring tools. Every change waits for a named person to approve it.
What if the agent gets it wrong?
Every finding links to the deploy, log line or metric behind it, so L2 can check it in seconds. Theories the agent rejected stay listed with the reason.
Can it run in our own cloud?
Yes, with bring your own cloud (BYOC). Your data stays in your environment.
Does our L1 team need production access?
No. The agent looks into the systems, and L1 and L2 see the findings in the ticket. Useful when engineering keeps production access tight.
Do agents replace our L1 and L2 teams?
No. Agents do the digging. Your teams decide, close tickets and talk to customers.