ServiceNow Autonomous Workforce: What's New in 2026?

For years, the phrase "AI in the enterprise" meant chatbots that answered FAQs, dashboards that surfaced recommendations, and automation that handled one step of a process before handing it back to a human. Useful, sure. But not exactly the transformation everyone was promised.

2026 is when that story started to shift at ServiceNow. At Knowledge 2026 in May, ServiceNow made a set of announcements around what it's calling the ServiceNow Autonomous Workforce, a framework of AI specialists that don't just assist employees, but actually carry out jobs from start to finish. Not partial jobs. Whole jobs.

So what does that actually mean, and what's new? Let's get into it.

What Is the ServiceNow Autonomous Workforce?

The ServiceNow Autonomous Workforce is ServiceNow's term for a new class of AI agents, called AI specialists, that operate with defined roles, authority, and governance inside your existing enterprise workflows. The idea is that these specialists don't just respond to requests; they interpret them, choose a course of action, execute across multiple systems, and document what they did.

ServiceNow president and chief product officer Amit Zavery put it plainly at Knowledge 2026: "Enterprises need AI that senses, decides, and securely acts." The distinction matters. Sensing and deciding are things AI has been doing for a while. Securely acting, within scope, with accountability, across real enterprise systems, is the part that's new.

The first specialist to reach general availability is the Level 1 Service Desk AI Specialist. It handles the most common IT requests, password resets, software access provisioning, network troubleshooting, around the clock, drawing on knowledge bases, historical incident data, and proactive remediation workflows. When it hits something outside its scope, it escalates to a human agent. When it resolves the issue itself, it logs everything.

More specialists are coming. ServiceNow has announced AI roles for IT operations, CRM, HR, finance, legal, procurement, and security and risk, each with domain-specific knowledge and process authority built in. This is what ServiceNow AI agents look like when they're designed for enterprise use rather than general-purpose assistance. Businesses planning to develop similar solutions can also learn how to choose the best AI agent development company to understand the expertise and capabilities needed for successful AI agent development.

The Numbers That Made People Pay Attention

The announcements at Knowledge 2026 came with some unusually specific results figures. These aren't projections, they're early outcomes from organisations already running the specialists:

• 99% faster IT case resolution - ServiceNow's own internal AI specialist vs human agents
• 90% of IT tickets targeted for autonomous resolution - Docusign's early results
• 98% employee request deflection rate - City of Raleigh, running the EmployeeWorks platform

Separately, ServiceNow's platform already resolves over 100 million customer cases monthly through its CRM automation layer, and 91% of AI-handled cases reach resolution without reassignment. These aren't pilot numbers. They're production scale.

ServiceNow EmployeeWorks: The Front Door for the Autonomous Workforce

One of the bigger announcements at Knowledge 2026 was ServiceNow EmployeeWorks, now generally available. It combines Moveworks' conversational AI and enterprise search (following ServiceNow's acquisition of Moveworks earlier in 2026) with ServiceNow's unified employee portal and workflow automation layer.

This gives you an AI-driven employee interface that operates in Microsoft Teams, Slack, or a web browser and does more than answering queries; it performs actions. What distinguishes it, according to Moveworks and ServiceNow GM Bhavin Shah: “An AI front door that doesn’t just summarize; it gets things done.”

In practice, this means that when an employee queries about their onboarding process, expense report approval, or IT access request, the EmployeeWorks takes care of the whole multi-system coordination to fix it, not just present the information. This knows the organization’s hierarchy, authorization processes, and it keeps a full audit log of all its actions. According to ServiceNow, the platform was designed for nearly 200 million employees around the globe.

The Governance Layer: AI Control Tower

Here's the thing that usually gets glossed over in coverage of ServiceNow AI automation: the part that makes autonomous AI safe for enterprise use isn't the agent capability. It's the governance layer.

ServiceNow's AI Control Tower is the platform's answer to that. As of the Knowledge 2026 announcements, all Control Tower capabilities are now built into every ServiceNow product and package, not an optional add-on, not an enterprise-tier upsell. Every ServiceNow customer gets it.

What it really does is detects AI agents operating throughout your environment, risk-ranks them, ensures least-privilege access (they’re limited to doing only what they're supposed to do, within their scope of operations), and gauges the business impact. Everything a ServiceNow autonomous AI expert does is auditible, and the policies are enforced at the workflow level, not as an afterthought to monitoring.

It is particularly relevant to regulated organizations. That is how you distinguish a governance-level implementation of AI from a proof-of-concept one. This is also the basis on which the principles of ServiceNow data governance and Workflow Data Fabric rest: policy-controlled flows of clean and traceable data behind any automation effort.

The Microsoft and Nvidia Partnerships: What They Mean in Practice

Two partnership announcements at Knowledge 2026 are worth paying attention to because they change the reach of ServiceNow AI agents significantly.

Microsoft

ServiceNow extended its AI Control Tower governance to cover Microsoft Agent 365, which means IT admins now have visibility and governance controls across AI agents running on both platforms in one place. Beyond governance, ServiceNow's AI specialists can now operate inside Microsoft 365 applications: Outlook, Word, PowerPoint. If an employee triggers a workflow through Teams or Outlook, the ServiceNow specialist can complete it without the employee ever switching context.

Nvidia

The expanded Nvidia partnership integrates Nvidia's accelerated computing infrastructure with the ServiceNow AI Platform. The practical effect: faster agent deployment at scale, which matters when you're rolling out AI specialists across tens of thousands of employees across multiple departments simultaneously.

What This Means for IT and ITSM Teams Specifically

For IT teams, the Level 1 Service Desk AI Specialist is the most immediately relevant announcement. Level 1 support, the repetitive, high-volume ticket queue that every IT team has, is exactly the kind of work the specialist is built for. Password resets, access requests, basic troubleshooting: tasks that consume significant human time but follow predictable patterns. The Docusign 90% autonomous resolution target and ServiceNow's own 99% speed improvement are Level 1 benchmarks, not enterprise-wide average figures.

What this doesn't mean is that ITSM teams disappear. The ServiceNow model is explicitly about freeing up human agents to handle the complex, judgment-heavy cases, not replacing them. The specialist escalates; a human makes the calls the specialist can't. For teams already running ServiceNow ITSM solutions, the autonomous workforce layer is an extension of what they've already built, not a rip-and-replace.

The longer-term consequence of this is one of capacity. If 90% of the tickets at Level 1 are solved through automation, how does the IT team utilize this capacity? The companies that will get ahead in this game will not wait for the specialist to fail but are already looking into redefining their role based on higher-value tasks such as problem management, architecture, and vendor governance.

Is Your Organisation Ready for the Autonomous Workforce?

What is the honest answer to this question? For the majority of enterprises, it will depend on the state of your ServiceNow environment right now. The autonomous specialists operate using your existing workflow, knowledge management systems, and process models. This means that if all of these are taken care of and governed properly, the specialists will be off to the races. Otherwise, they will inherit your legacy issues.

This is the implementation conversation that actually matters, not "which specialists should we enable?" but "is our data and process foundation in good enough shape for autonomous AI to work from?" The organisations getting early results like Docusign and the City of Raleigh aren't getting those numbers by chance. They have clean data foundations and well-defined process models to work with.

For those seeking solutions in how to get there, the ServiceNow App Engine offers the low-code environment for extending the workflows prior to any autonomous AI execution over them.If you're also considering the wider role of autonomous specialists in enterprise automation, exploring agentic AI development services can help you understand how AI agents can work across business processes and systems.

Dotsquares works with organisations at exactly this stage, assessing readiness, designing the data and governance foundations, and building the workflow layer that AI agent development runs on. If that's a conversation you're ready to have, it's a good time to start.

Frequently Asked Questions

What is the ServiceNow Autonomous Workforce?

The ServiceNow Autonomous Workforce is a framework of AI specialists, purpose-built AI agents, that carry out complete enterprise jobs from start to finish. Unlike traditional automation that handles single tasks, these specialists interpret requests, choose actions based on business context, execute across multiple systems, and maintain full audit trails. The first specialist (Level 1 Service Desk) reached general availability in June 2026, with HR, finance, legal, procurement, and security specialists to follow.

How are ServiceNow AI agents different from regular automation or bots?

Traditional bots follow fixed scripts; if the input matches a pattern, they take a predefined action. ServiceNow AI agents interpret context, make decisions based on business rules and historical data, and can handle requests that don't follow a single predictable path. They also operate with defined scope and authority; they know what they're allowed to do and escalate when they hit something outside that boundary, which is what makes them enterprise-grade rather than just technically capable.

What is ServiceNow EmployeeWorks?

ServiceNow EmployeeWorks is an AI-powered employee experience platform, now generally available, that combines Moveworks' conversational AI and enterprise search with ServiceNow's workflow and portal layer. Employees interact with it through Teams, Slack, or a browser, and it handles multi-system tasks, IT requests, HR queries, expense approvals, autonomously, without the employee needing to navigate multiple systems. It's designed to serve nearly 200 million employees globally.

How does ServiceNow govern its AI agents?

Through the AI Control Tower, which as of Knowledge 2026 is built into every ServiceNow product at no additional cost. It discovers AI agents running across your environment, risk-scores them, enforces least-privilege access (agents only act within their defined scope), and tracks every action with a full audit trail. Governance policies are enforced at the workflow layer, not just monitored after the fact, which is what makes it suitable for regulated and compliance-heavy environments.

What results are organisations seeing from the ServiceNow Autonomous Workforce?

Early results announced at Knowledge 2026 include: ServiceNow's internal AI specialist resolving IT service desk cases 99% faster than human agents; Docusign targeting autonomous resolution of 90% of IT tickets; and the City of Raleigh achieving a 98% deflection rate on employee requests. Separately, a global energy company cut threat containment time by 97% using the Autonomous Security and Risk specialist. These are early adopter figures, not average-case projections, but they indicate what's achievable on well-prepared implementations.