ServiceNow Enterprise AI Review

ServiceNow sells AI as part of its platform for IT service management (ITSM), HR service delivery (HRSD) and customer service management (CSM). The brand changed in 2026. Starting with Zurich Patch 12, Now Assist is ServiceNow Otto, and ServiceNow's release notes say product entitlements stay the same.

On May 5, 2026, ServiceNow announced Otto as one AI experience that combines Now Assist, Moveworks and AI Experience. Customers can use it first in ServiceNow EmployeeWorks and AI Control Tower, and ServiceNow says Otto will roll out across all products in the year ahead. Expect both names for a while: some current documentation pages carry ServiceNow Otto in the title and still mention Now Assist in the text.

How ServiceNow packages AI

ServiceNow structures its products and packages in three tiers:

• Foundation includes AI skills and routine AI agents, for task-based help such as pattern recognition and categorization.
• Advanced adds agentic workflows that automate entire steps in a process alongside the people doing the work.
• Prime adds what ServiceNow calls an autonomous AI workforce, described as fully independent AI specialists that make and execute decisions and run multiple agentic workflows.

AI skills come with every tier, across ITSM, CSM, HRSD and industry solutions. Administrators can configure ServiceNow's prebuilt agents at Foundation and Advanced, and building new agents from natural language instructions requires Prime. AI Control Tower is included in every tier with the same capabilities, and ServiceNow's announcement says it can log each AI interaction and enforce enterprise policies.

ServiceNow says the rollout of these tiers is independent of your upgrade cycle, so ask your account team which tier your current contract maps to and what it includes.

IT service management

ServiceNow Otto for ITSM lists generative AI skills for service desk agents with the itil role. Among them:

• Incident summarization, so an agent can understand an incident's context and answer the requester.
• Resolution notes generated on demand.
• Knowledge articles generated from incidents, to deflect similar issues later.
• Chat reply recommendations and email responses based on the incident.
• Draft emails to stakeholders during a major incident, built from an email template.
• Change request summaries, and an explanation of a change's risk assessment.
• Suggested steps generated from clusters of closed incidents with similar resolutions.
• Sentiment analysis of the requester.

Some of these skills are on by default, and an administrator can limit access to a subset of itil users with a role or group.

ServiceNow's machine learning features work alongside the generative skills. Predictive Intelligence offers three model frameworks in the current release. Classification sets a field such as incident category from the short description. Similarity recommends a resolution from a similar incident record. Clustering groups similar new incidents so a major outage shows up as a pattern.

HR service delivery

ServiceNow Otto for HRSD gives HR agents:

• Case summaries, and chat summaries that cover a handoff.
• Resolution notes generated when a case is changed to Closed Complete. The agent can accept or edit them before saving.
• Knowledge article drafts generated from a case, which the agent reviews before publishing.
• A sentiment score from -5 to +5 to help prioritize cases.
• Out-of-the-box filters that detect sensitive topics. ServiceNow notes that subjects such as workplace safety or employee compensation might not be suited to generative AI conversations.

For managers, ServiceNow documents agentic workflows that generate a personalized onboarding ramp-up plan for a new hire, which the manager reviews, revises and publishes through ServiceNow Otto in Virtual Agent, and an offboarding plan that identifies and transfers knowledge from departing employees. Out-of-the-box agentic workflows and AI agent records are read only. To change one, you duplicate it first.

Customer service

ServiceNow Otto for CSM covers case and chat summarization, resolution notes, knowledge generation, call summarization and email responses. The CSM AI Agent Collection adds prebuilt, configured AI agents that complete tasks within customer service workflows. ServiceNow's documentation tells customers to keep human oversight of the output and not to rely only on AI-generated output for decisions.

AI agents and models

AI Agent Studio is where teams create, manage and test AI agents and agentic workflows. An orchestrator coordinates the agents, tools give them the capabilities to complete their tasks, triggers start a workflow, and the output appears in the ServiceNow Otto panel.

ServiceNow lists its own Now LLM Service alongside Azure OpenAI, Google Gemini and Anthropic's Claude through AWS as model providers, and you can bring your own LLM provider. Some out-of-the-box skills, agents and agentic workflows don't support every provider, so check the ones you plan to use against the provider your security team approves, along with the data routing for your region.

Strengths and trade-offs

What ServiceNow's documentation shows in its favor:

• The AI works on the incidents, cases and knowledge articles already in the platform.
• Several skills build in a review step, such as accepting or editing resolution notes and reviewing knowledge drafts before they publish.
• AI Control Tower comes with every tier.

What to weigh against that:

• The tiers are new and the brand is changing, so contracts, documentation and in-product labels may not match for a while.
• Building new agents from natural language requires Prime.
• Prebuilt agentic workflows are read only until duplicated, and each duplicate becomes something your team maintains.
• No published prices, so every comparison depends on a quote.

Running a pilot

1. Pick one queue in ITSM, HRSD or CSM, and two or three skills your tier already includes, such as incident or case summarization and resolution notes.
2. Record how that queue works today: how long a handoff takes to understand, and how often resolution notes are thin or missing.
3. Turn the skills on for a small group of agents, and have them accept, edit or reject each output.
4. Track how often agents edit the output, and read a sample of outputs each week.
5. Confirm the model provider and data routing with your security team before widening access.

For how a pilot like this fits a wider plan, see AI for Enterprise: The Practitioner's Strategic Guide.

Pricing

ServiceNow does not publish prices. Its product tiers page directs customers to their ServiceNow account team for availability and entitlement details. Ask for a quote broken out by product (ITSM, HRSD, CSM) and tier, which AI skills and agents each tier includes, whether the agents or workflows you want require Prime, and which model providers are available in your region.

The details above match ServiceNow's documentation and newsroom as of November 1, 2026.