ServiceNow tackles ‘sidecar AI’ chaos with agentic workforce strategy

ServiceNow is betting its platform can save enterprises from repeating the mistakes of the past, warning that a proliferation of disconnected “sidecar” artificial intelligence (AI) tools is creating a wave of complexity that can stifle business transformation.

Speaking at the ServiceNow World Forum in Sydney today, company executives unveiled a strategy centred on agentic workforce management, designed to orchestrate teams of AI agents that work autonomously across the enterprise, all governed from a single platform.

Colin Fleming, ServiceNow’s chief marketing officer, argued that while AI represents the “defining shift of our time”, most businesses are failing to capitalise on its potential. He pointed to the company’s research, which showed that business AI maturity dropped by 20% last year, despite the hype.

“We are all repeating the same mistake that we made with software,” Fleming told the audience of business and technology leaders. He cited the average company’s use of 367 disconnected applications, a “hornets’ nest of complexity” that businesses are now replicating with AI.

“Each and every one of those 367 different applications now wants its own AI bolted onto the side like a sidecar,” said Fleming. “None of them are connected, none of them knows what each other is doing, and none of them actually moves the work forward. They talk, they summarise, but they don’t do the work.”

ServiceNow’s answer is to move beyond task-oriented bots to what it calls an agentic workforce – a collective of orchestrated AI agents that can handle complex workflows. This is managed through tools like the AI Agent Orchestrator, which coordinates agents, and the AI Control Tower, which provides a central command centre for governance, security and performance tracking.

Gaurav Rewari, senior vice-president and general manager of data and analytics products at ServiceNow, however, noted that “the road to an AI agentic heaven often goes through a data hell with data silos, fragmented systems, missing context in the data analytics and performance bottlenecks”.

To address those challenges, Rewari pointed to ServiceNow’s efforts to create an AI-ready data foundation. These include RaptorDB, the company’s proprietary database built for both transactional and analytical workloads, and Workflow Data Fabric, which enables zero-copy integration with data platforms like Snowflake, Databricks and Google BigQuery.

In addition, Rewari said ServiceNow’s recent acquisition of Data.world, a data governance company, will help to enrich data with semantic meaning and improve governance across distributed data landscapes. “This makes your AI agents smarter, so they can make better decisions, take better actions and ultimately achieve better results,” he added.

Australian mining and infrastructure giant Orica also took to the stage to share its experience with ServiceNow. The company is an early adopter of ServiceNow’s Now Assist AI copilot, and has deployed over 20 AI skills and five agentic AI use cases.

Bradley Hunt, manager of DevOps and regional apps at Orica, said the company has deflected 94% of service requests away from human service desks, along with a 70% increase in AI usage among its IT staff over the past six months. “We now have 75% of IT agents who are assigned cases using AI,” said Hunt. “It resulted in about 12,000 AI actions every month.”

Rachael Sandel, Orica’s group CIO, noted that success with AI was not just about technology, but about strategy and governance. “One of the key things that really made a difference for us was establishing strong governance and securing executive alignment from the outset,” she said.

Sandel said Orica treats AI as a core part of its strategy, managed by a cross-functional centre of excellence. “This top-down support and the transparency we’re bringing brings clarity and confidence to our teams to move fast while staying safe, responsible and focused on value,” she added.

ServiceNow itself is also seeing significant gains from using its own agentic workforce, automating 97% of software provisioning requests and autonomously resolving 85% of routine internal IT support issues.

“Enterprises aren’t asking if they’ll adopt agentic-enabled ways of working anymore; they’re wrestling with how to frame it responsibly,” said Amy Loomis, research vice-president at IDC. “With trust and governance just as critical as productivity, ServiceNow’s integrated approach positions it as a leading voice in defining how organisations manage the relationship between workers and AI agents at scale.”