ServiceNow is dealing with 90% of its personal worker IT requests autonomously, resolving circumstances 99% sooner than human brokers. On Thursday it introduced the product know-how it needs to make use of to do the identical for everybody else.
Organizations have spent three years operating pilots that stall when AI will get to the execution layer. The agent can establish the issue and suggest a repair, then hand it again to a human as a result of it lacks the permissions to complete the job or as a result of nobody trusts it to behave autonomously inside a ruled setting.
The hole most groups are hitting isn't functionality. It's governance and workflow continuity.
ServiceNow's reply is a brand new framework referred to as Autonomous Workforce; a brand new employee-facing product referred to as EmployeeWorks constructed on its December acquisition of Moveworks; and an underlying architectural method it calls "position automation."
From ticketing system to AI workforce
ServiceNow has been constructing towards this for twenty years. The platform began as a ticketing system, advanced right into a workflow automation engine, and spent the final two years layering AI onto that basis by means of its Now Help product.
What's completely different is that the brand new method stops treating AI as a characteristic sitting on high of workflows and begins treating it as a employee working inside them. That shift, from AI that assists to AI that executes, is the place the broader enterprise market is headed. ServiceNow is making a selected architectural guess about the right way to get there.
The announcement has three components: ServiceNow EmployeeWorks lets workers describe an issue in plain language and have it mounted with out submitting a ticket; Autonomous Workforce executes work finish to finish; and position automation is the architectural layer that governs how these specialists function inside present enterprise permissions.
Most enterprise AI assistants together with Microsoft Copilot and Google Gemini require workers to know which software handles which drawback. Moveworks, which had 5.5 million enterprise customers earlier than the December acquisition, was constructed round a single entry level that routes throughout that ambiguity mechanically.
Bhavin Shah, founding father of Moveworks and now SVP at ServiceNow following the acquisition, framed the issue instantly in a briefing with press and analysts.
"Over the past two years, organizations have raced to undertake AI, however in lots of circumstances that rush has created fragmented instruments, disconnected AI experiences and workers bouncing between programs simply to get easy issues performed," he stated.
Why position automation is completely different from an everyday agent
ServiceNow is proposing a brand new architectural layer it calls position automation, and it differs from the brokers most enterprises are already operating.
Standard AI brokers are task-oriented: they're given a purpose, they cause towards it and in doing so that they determine what they're allowed to do at runtime. That creates issues in enterprise environments the place governance, audit trails and permission boundaries aren't optionally available.
With role automation, an AI specialist doesn’t cause its method into permissions. It inherits them. The identical entry management framework, CMDB(configuration administration database) context, SLA (service stage settlement) logic and entitlement guidelines that govern human employees on the ServiceNow platform govern the AI specialist from the second it’s deployed. It can’t exceed its outlined scope. It can’t self-escalate privileges primarily based on what it learns mid-task.
The corporate attracts a three-tier distinction: job brokers deal with particular person automation steps, agentic workflows combine deterministic and probabilistic execution, and position automation sits above each as a completely virtualized worker position with outlined tasks and pre-inherited governance.
The primary product constructed on this structure, the Stage 1 Service Desk AI Specialist, handles widespread IT requests finish to finish — password resets, software program entry provisioning and community troubleshooting — documenting every decision and escalating to a human agent solely when it hits one thing outdoors its outlined scope.
'Don't chase butterflies'
Alan Rosa has seen what occurs when AI governance fails in healthcare. As CISO and SVP of infrastructure and operations at CVS Well being, he manages AI deployment throughout 300,000 workers the place compliance isn't optionally available.
Talking on the similar briefing, his framework for scaling AI maps instantly onto what ServiceNow is claiming architecturally. CVS Well being was already a buyer of each ServiceNow and Moveworks earlier than the December acquisition. Rosa stated the mixture of the 2 platforms is encouraging and that the potential is "coming to life," although CVS Well being has not dedicated publicly to deploying Autonomous Workforce.
"Boring is gorgeous," Rosa stated. "Predictable. Secure. You must begin with accountable, explainable AI. No bias, no hallucinations, clear guardrails. Everybody understands the foundations."
On the temptation to chase the latest AI capabilities earlier than governance is in place, he was direct: "Don't chase butterflies. Concentrate on gritty, unsexy, operational use circumstances. Those with actual ROI that have an effect on individuals's lives."
Rosa's method treats AI as a constantly evolving set of capabilities requiring dynamic somewhat than static testing. CVS Well being runs each AI use case by means of medical, authorized, privateness and safety assessment earlier than it touches manufacturing.
"Static assessment doesn't reduce it when AI is studying and adapting," he stated. "Wash, rinse, repeat."
Rosa's framework requires governance to be embedded within the deployment structure from the beginning, not retrofitted after an issue surfaces. That’s exactly the declare ServiceNow is making about position automation. AI specialists that inherit present enterprise permissions and workflow logic are structurally much less prone to break governance boundaries than brokers that decide their very own scope at runtime.
What this implies for enterprises
For any group evaluating agentic AI, no matter vendor, the sensible query is easy: Does your AI governance reside inside your execution layer, or is it sitting on high of it as a coverage doc that brokers can cause previous?
That’s what ServiceNow is attempting to resolve with Autonomous Workforce and EmployeeWorks, baking governance and workflow context instantly into the agentic layer somewhat than bolting it on afterward. For practitioners, the start line is governance structure, not functionality. Earlier than deploying any agentic AI, map the place your permissions, workflow logic and audit necessities truly reside. If that basis isn't in place, no agent framework will maintain at enterprise scale.
"Scale and belief go collectively," Rosa stated. "For those who lose belief, you lose the best to scale."
