ServiceNow has unveiled an innovative AI solution designed to assist human employees, with plans to introduce its Autonomous Workforce later this year, potentially stepping into human roles.
Next quarter, ServiceNow intends to launch the inaugural component of its Autonomous Workforce: the Level 1 Service Desk AI specialist.
This AI agent is set to independently identify and fix routine IT support issues, including password resets, software access grants, and network problems. Its operations will be informed by corporate knowledge bases, past incident records, and established procedures. Operating round-the-clock, this agent will allow human staff to concentrate on higher-level strategic initiatives by handling routine duties within the established enterprise scope, authority, and governance framework, as reported by the company.
Internally, ServiceNow is already utilizing this agent, reporting that it manages over 90% of employee queries. Furthermore, it completes these tasks nearly twice as quickly as its human counterparts, all while adhering to the critical business context and governance standards expected in an enterprise environment.
ServiceNow’s AI specialists, such as the Level 1 Service Desk agent, are engineered to collaborate with human employees. They function within a meticulously defined operational scope, which is governed by the identical permissions granted to a human fulfilling that specific role.
“By design, AI specialists are prevented from surpassing their designated authority or autonomously elevating their permissions in memory, irrespective of the reasoning outcomes from the initial stage of the AI-driven decision and execution process,” explained John Aisien, SVP central product management at ServiceNow, during a press briefing. “Conversely, these AI specialists base their determinations on current enterprise data, leveraging real-time insights regarding assets, access rights, ownership, up-to-the-minute permissions, and established resolution protocols via our enterprise data foundation and context graph.”
ServiceNow stated that by integrating probabilistic intelligence with structured workflow orchestration, its AI specialists are capable of understanding requests, applying business context to identify the most appropriate action, and executing it under the supervision of ServiceNow’s AI Control Tower. Subsequently, they inform the relevant employee and update the knowledge base. Should an issue remain unresolved, it is then escalated to a human agent at Level 2 or Level 3 for deeper inquiry.
Introducing EmployeeWorks
Additionally, ServiceNow revealed EmployeeWorks, now accessible, describing it as “a conversational gateway to the enterprise.” This solution functions as a personalized assistant, integrating conversational AI and enterprise search capabilities from Moveworks, a company ServiceNow recently acquired, alongside ServiceNow’s proprietary unified portal and autonomous workflows, according to Bhavin Shah, Moveworks’ founding CEO and current general manager for Moveworks and AI at ServiceNow.
“Workers will no longer need to determine which agent to activate, where to navigate, or ponder ‘should I use this platform or that one?’” he explained. “It simply operates seamlessly.” He further noted that the service incorporates protocols like MCP and A2A, facilitating “secure, scalable interaction among agents and various business systems.”
ServiceNow confirmed that EmployeeWorks comprehends organizational hierarchies, approval processes, and authorization rules, allowing it to perform tasks necessitating coordination across multiple systems, all while preserving governance and audit trails. For instance, it can extract data from a SharePoint document, cross-reference a Slack discussion, synthesize the information to initiate an action, or manage approvals, orchestrate complex workflows, and update various systems, consistently adhering to enterprise policies.
Shah highlighted that EmployeeWorks operates independently of vendor ecosystems, can address employee inquiries without requiring them to transition between different applications, and offers immediate integration capabilities along with comprehensive enterprise search functionality.
Concerns Regarding Automation Initiatives
While analysts generally endorse ServiceNow’s strategic trajectory, they express some concerns regarding the recent product announcements.
The integrated governance features of Moveworks appear “impressive,” noted Info-Tech Research Group Advisory Fellow Scott Bickley. However, deploying EmployeeWorks will necessitate substantial preliminary work, encompassing the documentation of workflows, the refreshment of knowledge bases, data cleansing, and the precise definition of approval routes, complete with safeguards and contingencies to account for all potential scenarios.
Concurring with the sentiment, Greyhound Research Chief Analyst Sanchit Vir Gogia remarked, “ServiceNow is taking the correct path by integrating AI deeply into workflow control.” He added, however, that “a sound direction doesn’t automatically equate to mature execution. This strategy’s real-world viability will be assessed in complex, heavily regulated, and multi-system operational settings, rather than in oversimplified service desk scenarios.”
According to Gogia, this represents a significant architectural shift: “AI is now being structured to act as a delegated entity within specific job functions, which fundamentally alters the landscape of accountability.” He elaborated, “This explains why ServiceNow’s focus on deterministic workflow orchestration aligns perfectly with enterprise requirements. While AI models inherently operate on probabilities, businesses demand results that are reliable, traceable, and adhere to established policies.”
However, ServiceNow omitted clarification on accountability should an AI specialist deviate from its intended operation.
Melody Brue, Principal Analyst at Moor Insights & Strategy, voiced a concern that “AI agents might emerge as a novel layer, potentially bypassing numerous applications currently in use.” She clarified that “ServiceNow’s objective is to position itself above this, orchestrating agents and workflows across various systems, rather than merely introducing another tool that could eventually be superseded.”
She further asserted that AI’s role has evolved beyond simply delivering marginal efficiency gains. Presently, “it is imperative for AI to aid in extracting untapped value from enterprise data and workflows. By embedding AI into core systems of record and structured workflows, ServiceNow intends to transition from generating passive reports to deploying agents that actively leverage insights.”
Gogia presumes that businesses will indeed embrace autonomous AI. The crucial inquiry, he posited, revolves around whether they can effectively manage it without compromising operational integrity.
Bickley raised another issue: the financial implications for businesses. SaaS providers employ diverse usage-based “AI credit” systems for their AI services, making it challenging to precisely forecast and budget for AI credit consumption, he explained.
“A transparent migration path for converting traditional seat subscriptions into AI credits is essential,” Bickley asserted. He emphasized that “CFOs will find a fluctuating pricing model that undermines budgetary foresight unacceptable. This critical concern appears to be overlooked by ServiceNow, and indeed, the wider SaaS industry, as they intensify their ambitious AI deployment strategies.”
Originally, this piece was published on CIO.com.
