---
key: check-how-novaai-for-calls-is-running
title: Check how NovaAI for calls is running
audience: staff
section: communications/calls
routes: ["/communications/nova-ai"]
permissions: ["communications.call_intelligence.read"]
features: ["communications.call_intelligence"]
since: 0.0.1
updated: 0.0.1
status: published
summary: Use NovaAI Ops to see whether NovaAI for calls is ready, what it has cost and how busy it has been over the last 30 days, for the whole organization or one phone line.
keywords: [novaai status, ai cost, ai spend, call ai health, is ai working, novaai ops, ai backlog, ai failures, emergency stop]
related: [turn-on-novaai-for-calls, review-a-call-with-novaai, set-up-a-phone-line]
---

NovaAI Ops is not in the sidebar. In the sidebar, open **Settings**, then choose **NovaAI for Calls** under **Communications**, and press **NovaAI Ops**. **Live Calls** also has **Open NovaAI ops**. The page address is `/communications/nova-ai`. The page title reads **NovaAI Ops**.

## Before you start

- NovaAI for calls has to be on for your organization; see [[turn-on-novaai-for-calls]]. Until it is, the page reads **This page is not available for this organization.**
- Your role needs permission to read NovaAI results.
- This page only shows figures. Nothing here changes a setting; **Open configuration** takes you to the settings.

## Steps

1. Under **Observable scope**, leave **Runtime scope** on **Organization observable scope** to see the whole organization, or pick one phone line.
2. Under **Operating console**, read **Deployment readiness**. The chip beside it says whether NovaAI is ready to run.
3. Read **First-live operator actions**. Each item names the part that has a problem and, when Nova knows it, the next step to fix it.
4. Read the tiles for the last 30 days: **Provider invocations**, **Provider failures**, **Total spend**, **Average latency**, **Backlog** and **Active concurrency**.
5. With a phone line picked, press **Manage this line override** to change NovaAI for that line only.

## What you see

- **Provider invocations** counts the requests NovaAI sent to its AI service, and **Provider failures** how many of those failed. **Total spend** is what they cost. **Backlog** counts work still waiting its turn.
- **Rollout and canary** shows whether shadow mode is on and which pilot rule, if any, matched the scope you picked.
- **Kill switch posture** lists any emergency stops turned on in settings, or reads **No active kill switches for this scope.**
- **Realtime reasons** and **Post-call reasons** explain why NovaAI held back during or after calls.
- **Provider and telephony config** and **Workers, beat, and storage** are technical checks. If one shows a problem, pass it to whoever set up Nova or to Nova support.

## If something looks different

- The page reads **Unable to load NovaAI operations.**: reload the page. If it keeps happening, contact Nova support.
- A notice reads **Runtime line inventory is unavailable. Organization-wide observability will continue to load when possible.**: the list of phone lines could not load. The organization-wide figures still show.
- The phone line you picked goes back to the whole organization: you no longer have access to that line.
- **Manage this line override** is missing: pick a phone line under **Runtime scope** first.

[[route:/communications/nova-ai|Open NovaAI Ops]]