If you have been following ServiceNow's AI journey, you may have noticed quite a few changes recently .One of the biggest ones is the move from Now Assist to ServiceNow Otto. At the same time, Now Assist Center has evolved into AI Admin Center.

So, if you are a ServiceNow AI administrator, there is a natural question:

With all these AI capabilities coming into the platform, where do I actually manage them?

This is where AI Admin Center comes in.

ServiceNow describes AI Admin Center as a single control hub that brings together AI capabilities and configuration functions so administrators can set up and manage AI from a unified experience.

Let's have a look at what this means from an AI administrator's perspective.

What is AI Admin Center?

If you have worked with Now Assist (Otto) before, you probably know that AI configuration can involve multiple applications, plugins, skills, settings, and consoles.

As the number of AI capabilities grows, this can quickly become difficult to manage. AI Admin Center brings many of these activities together.

Think of it as the control center for AI administration in your ServiceNow instance.

Depending on your instance, release, applications, and entitlements, the capabilities you see can vary.

And this is something worth keeping in mind throughout this blog.

AI Admin Center is not just about turning AI on. It is about understanding what AI capabilities are available, configuring them, monitoring them, and making sure they are actually delivering value.

1. Start with AI Readiness

Before we start enabling AI capabilities, there is one question we should ask:

Are we ready for AI?

AI Admin Center provides an AI readiness experience that helps AI administrators/stakeholders identify areas that may need attention before adopting AI.

This is important because AI readiness is not simply about installing an application. Your data, configuration, security, and existing processes all have an impact on the outcome.

For example, let's say we want to use Otto to summarize incidents.

If the incident data is incomplete or inconsistent, Otto can still generate a summary. But that doesn't necessarily mean the summary will be useful.

This is something I always like to emphasize:

AI can work with poor data. It doesn't mean the outcome will be good.

So, before enabling AI, understand where the gaps are and address them.

2. Discover and manage AI capabilities

Once we know that the instance is ready, the next question is:

What should we actually enable?

This is where AI Admin Center becomes useful. The platform provides visibility into the AI capabilities available to the instance and helps AI administrators configure and activate them.

And with the move from Now Assist to ServiceNow Otto , you will increasingly see Otto terminology across these capabilities. ServiceNow's current release documentation refers to Otto as its new AI experience brand.

But I would not recommend enabling everything just because it is available. Start with the business problem.

For example:

"Our service desk agents spend too much time reading long incident histories."

That could be a good candidate for summarization. Or:

"Our agents spend too much time writing repetitive responses."

That could be a good candidate for generative assistance. The question should not be:

"What AI features do we have?"

It should be:

"Where can AI make a meaningful difference?"

3. Otto skills and AI Agents

This is another area where terminology can get confusing. We now have Otto capabilities/skills as well as AI Agents.

They are not the same thing. A skill generally helps a user perform a specific task using AI.

An AI Agent can go further. It can work toward a goal, reason through the task, use tools, and perform multiple actions.

For example, instead of simply telling a user how to troubleshoot an issue, an agent could potentially:

Understand the request.
Gather relevant information.
Decide what needs to be done.
Use the required tools.
Perform the action.
Verify the result.

This is where AI Agent Studio becomes important.

I find it useful to think about the different components like this:

Capability	What it is mainly used for
AI Admin Center	Administer and monitor AI
AI Agent Studio	Build and configure AI agents
AI Control Tower	Govern and oversee AI
ServiceNow Otto	Deliver AI experiences and capabilities

This makes the overall picture a little easier to understand.

4. Where will users experience AI?

Configuring an AI capability is only one part of the story.

The next question is:

Where will the user actually experience it?

AI can be surfaced in different ServiceNow experiences, including workspaces, Virtual Agent, search, record experiences, and conversational experiences.

This is important for adoption.

If users have to leave the application they are working in to use AI, it may not become part of their normal workflow. But if AI is available exactly where the work is happening, it becomes much more useful.

For example, an agent working on an incident may benefit from AI assistance directly within the incident experience instead of opening another application.

That is where the combination of Otto + the ServiceNow user experience becomes interesting.

5. Monitor usage and adoption

Now comes one of the questions I consider extremely important:

Are people actually using the AI capabilities we enabled?

Because enabling AI is easy. Getting users to adopt it-and proving that it is providing value-is much harder.

AI Admin Center provides monitoring capabilities to help AI administrators understand AI usage, performance, and adoption. The current AI Admin Center experience also brings monitoring of skills, AI agents, and assistants into a more consolidated experience.

This allows us to move beyond:

"We have enabled AI."

and start asking:

"What is actually happening after we enabled it?"

Which capabilities are being used? Which ones aren't? Are users getting value from them? Are agents completing their tasks successfully?

These are much more useful questions. And one point I would strongly recommend remembering:

Activation is not success.

Even adoption alone is not necessarily success. We ultimately need to understand the business outcome.

6. AI governance cannot be an afterthought

Of course, AI administration is not only about configuration. It is also about control.

As we introduce generative AI and AI Agents, we need to think about:

Who can use the capability?
What data can it access?
What actions can an agent perform?
What permissions does it have?
What guardrails are in place?
Where should human approval be required?

This becomes even more important with agents. An agent may not simply generate an answer. It may actually perform an action.

So when designing an agent, I would ask two questions:

What should this agent be able to do?

and

What should this agent never be allowed to do?

The second question is just as important as the first one.

7. Finding automation opportunities

Another area I find particularly interesting is automation opportunity discovery. Most organizations have plenty of processes that could potentially be automated.

The challenge is identifying the right ones. AI Admin Center can help administrators identify potential automation opportunities, including opportunities where AI agents could provide value.

This changes the conversation. Instead of starting with:

"I want to build an AI agent. What should it do?"

we can start with:

"Here is what is happening in my instance. Where could AI help?"

I think that is a much better starting point. But even when the platform identifies an opportunity, we still need to validate it.

Is the process well understood? Is the required data available? Are the actions safe? Can we measure the outcome? And most importantly:

Does this really need an AI agent?

Sometimes a Flow is still the better solution. Not every automation problem needs AI.

8. AI helping the administrator

This is probably the part I find most interesting. We started by asking:

How can AI help our users?

Now we are starting to ask:

How can AI help the administrator?

AI Admin Center is moving in this direction with capabilities around readiness, recommendations, conversational administration, and AI-assisted administration.

That changes the role of the administrator.

Instead of spending all our time finding configuration, checking dependencies, and troubleshooting settings, we can increasingly focus on:

Architecture
Governance
Data quality
Adoption
Business value
AI strategy

We are moving from Administrator manages AI toward Administrator works with AI to manage AI. And I think that is going to be an important change.

My approach to AI administration

If I were starting an AI implementation today, I would keep the approach fairly simple.

1. Start with the problem

Don't start with the AI feature. Start with what users are struggling with.

2. Check readiness

Understand whether the platform, data, and processes are ready.

3. Pick the right capability

It could be Otto, an AI Agent, automation, search-or sometimes no AI at all.

4. Start small

Pick a use case where the outcome can be measured.

5. Test properly

Don't test only the happy path.

6. Think about security

Review data, roles, permissions, tools, and guardrails.

7. Monitor

Understand adoption and performance.

8. Measure value

Ask whether the capability is actually saving time, improving resolution, or reducing manual effort.

9. Expand

Once you have proven one use case, move to the next.

Final thoughts

For me, the interesting thing about AI Admin Center is not simply that it gives us another place to configure AI. It represents a change in how we think about ServiceNow administration.

Earlier, our focus was primarily on configuring the platform.

With AI, we now need to think about an additional lifecycle:

Readiness → Configuration → Deployment → Adoption → Monitoring → Governance → Optimization

And increasingly, AI itself is becoming part of that lifecycle. We can use AI to identify opportunities. We can use AI to help with administration. We can use AI Agents to perform work.

And we can use AI Admin Center to help us understand and manage the AI capabilities running in our environment.

So, perhaps the real question for a ServiceNow administrator is no longer:

"How do I configure AI?"

It is:

"Where should we use AI, how do we use it safely, and are we actually getting value from it?"

And that, for me, is where AI Admin Center fits into the bigger ServiceNow AI story.

News Jace Pro NJP

ServiceNow AI Admin Center: Managing the New Era of Otto and AI