Interested in a ServiceNow event built for developers? Registration for now[dev]26 is officially open!

Sharon_Barnes
ServiceNow Employee

That Otto Panel in AI Admin Center Can Do a Lot More Than You Think

Requirements

Family Release Australia P4
Zurich P11
AI Admin Center 5.1.9
Role (admin) sn_na_center.nac_admin

Overview

💡 Rename note: AI Admin Center (AIAC) was formerly known as Now Assist Center (NAC). Everything in this post applies to both. The app ID hasn't changed: sn_na_center.

You might not realize it, but the Otto panel pinned to the right side of your browser is more helpful than it seems. It's been a part of every page since Australia, and hidden behind it is a whole set of specialized AI agents ready to assist you with managing AI requests, tailored to what you're asking.

Once you understand what's there, it changes how you work in the console. Instead of hunting through menus to configure something, or opening a new tab to search the docs, you just describe what you want. The panel can handle the rest.

Let me walk you through it.

What the Conversational Interface Covers

Before getting into the individual agents, here's what the Otto panel can do at the top level. When you chat with Otto, you can:

  • Kick off guided setup workflows for generative AI skills. The companion either takes you to the setup page or executes steps on your behalf, but always asks for confirmation before anything actually changes.
  • Activate or deactivate specific generative AI skills. You'll see skill details and current state before you confirm anything.
  • Query AI configuration details about your instance.
  • Ask general questions about AI features, concepts, and admin tasks. Responses come with references and links to source documentation. You don't have to leave the console to get answers anymore.

Three Agents Doing the Actual Work

When you send a message, the Otto panel routes it to one of three specialized agents depending on what you're asking. They can also work together on the same request. The routing is automatic and you'll see which agent responded in the reply.

Self-Healing AI Agent

This one is for when something is broken and you're not sure why. Describe the problem in plain language, and the agent matches your description against a library of known AI administration failure patterns. It returns a root-cause hypothesis and a confidence level, so you can see how certain it is before you do anything.

Behind the scenes, two sub-agents work together. The SHA Triage Agent goes first and checks whether your problem description has enough to work with. If it needs more detail, it'll ask one clarifying question (max two turns). Once it has what it needs, the SHA Diagnostic Agent runs the analysis and gives you an actionable result.

The agent holds context throughout your session, so you can ask follow-ups without re-explaining everything. That alone saves a lot of back-and-forth.

It's on by default. You need the sn_na_center.nac_admin role. The kinds of things it can dig into:

  • Configuration or setup problems in AI Admin Center
  • Plugin installation or activation failures
  • Missing or disabled AI features
  • Role or permission misconfigurations affecting AI access
  • Compatibility conflicts between plugins or versions

Depending on what it finds, it might walk you through specific configuration steps, surface a relevant KB article, suggest settings to verify, or let you know it's time to escalate to support.

AI Admin Center Help AI Agent

This one handles the "how do I" and "what is" questions. Ask it something that sounds like a product question rather than a problem, and it searches ServiceNow documentation and training resources and comes back with a real answer, including references and links so you can go deeper.

It also holds conversation context, so follow-up questions work the way you'd expect. Role required: sn_na_center.nac_user.

AI Analytics Q&A Agent

This one is scoped specifically to questions about the AI analytics dashboards in AI Admin Center. What does this KPI mean, how is this widget calculated, what changed between releases. That kind of thing. So far, it is an expert on skills and assistant metrics.

Role required: sn_na_analytics.viewer. What it can answer:

  • KPI definitions in AI analytics dashboards (CSAT, deflection rate, etc.)
  • How a specific widget or metric is calculated
  • Data flows and processing context for AI analytics
  • Deflection log states and common scenarios
  • What changed in the dashboards between releases

Just Try It 🚀

Honestly the fastest way to get a feel for this is to open AI Admin Center, pull up the Otto panel, and either describe a configuration issue you've been sitting on or ask a question about one of your dashboards. The routing between agents is automatic, and from there it's pretty intuitive.

If you want to go deeper on any of these agents, the docs are worth a look:

Check out the Now Assist Center Hub for more resources