ServiceNow makes it clear that AI agents are no longer future music, but shows how these ‘digital twins’ are already performing real tasks in various organizations today.
Under refreshing air conditioning and in the company of thousands of customers and interested parties, ServiceNow’s annual Knowledge conference took place. The company showed that it is clearly playing the AI card: AI should form the foundation of every work process.
CEO Bill McDermott, a vessel full of charismatic one-liners with his trademark sunglasses, kicked off the conference with a clear message: “We’re not building demos, we’re building an operational AI platform.” The tone was set. And that platform got a name this week: the ServiceNow AI Platform, with AI Agent Fabric and the AI Control Tower as support.
Digital Twins
The showpiece was the AI Agent Fabric, an underlying architecture that allows AI agents to function as digital colleagues. Amanda Joslin, Senior Director AI Agents, called it “more than just a chatbot”: agents actually perform actions based on data and previous interactions.
To keep this on track, there’s the AI Control Tower. This dashboard displays all active agents, but also what they’re doing, whether they’re meeting certain KPIs, and where adjustments might be needed. According to Amit Zavery, COO at ServiceNow, this kind of overview is essential. “Without a central dashboard, you sow chaos. The Control Tower acts as “a bridge between freedom and responsibility”.
It’s a place where both employees and customers can see together what agents are doing and why. The tool works not only with ServiceNow’s own agents but also with third-party ones. “You can view and use all agents from one place,” says Zavery.

ServiceNow’s ambition goes beyond just automation. McDermott dreams of an ‘AI twin’ for every employee. “Imagine: an AI agent that shows your best opportunities, drafts your proposals, calculates your compensation. All in milliseconds.”
During his keynote, McDermott also warned about the cost of fragmented software landscapes. “Forty percent of office productivity is lost because people click back and forth between seventeen different applications per day,” he said. “That’s pure madness and a productivity disaster.”
This destruction of time, as he called it, is exactly what ServiceNow aims to prevent with one platform. No more fragmentation between tools, but a coordinated approach where AI and data come together.
AI with Memory and Do-It-Yourself
Long-term memory in AI is not a first in itself. Yet, the way ServiceNow connects that memory layer with business data is unique. The Workflow Data Fabric ensures that agents have access to live business data without having to copy it first. For example: An HR agent who knows that someone submitted a request yesterday will take the right follow-up step today. And that without anyone having to follow up manually.
“AI is not the next step, it’s the biggest opportunity of this century. Bigger than internet, mobile, and cloud combined. Our only chance for progress and survival.”
Bill McDermott, CEO ServiceNow
With the AI Agent Studio, users can build their own agents via text prompts, without programming knowledge. A live demo showed how to create an agent in less than five minutes that locates, summarizes, and forwards documents. The combination of the user-friendly interface, the use of business data, and control via Control Tower should, according to ServiceNow, provide a valuable work experience.
Who’s already using the AI platform?
Customers shared their experiences in breakout sessions and during keynotes. At Astra Zeneca, agents are deployed for onboarding, HR administration, and compliance. “We save time, but more importantly: we avoid mistakes,” said an HR director. “And the experience is more consistent than when ten people handle the same process differently.”
Aptiv uses agents in supply chain workflows. “Sometimes an order comes in three times from the same supplier, or a critical inventory is reported late. With agents, we avoid that.” At UKG, agents are being tested for answering personnel questions, which otherwise would have to go manually to HR.
Applications today are mainly focused on predictable processes such as repeated questions, approvals, or document management. More complex domains like risk management or customer interaction are still in the works. ServiceNow acknowledged this as well.
CRM needs renewal
“What you often see,” said Vice President Product Management Michael Ramsey, “is that companies think the customer experience stops at the interface. A beautiful website or a good conversation. But if something gets stuck internally, you still get frustration from both the customer and the employee(s). That’s the messy middle.”
That’s the collection of all internal processes that customers normally don’t see, but do feel when they go wrong: manual approvals, double data entry, missing information between departments. ServiceNow wants to solve this with an AI-native CRM approach that aligns data and processes.
read also
ServiceNow Launches New AI Platform for All Business Departments
Zavery was clear: “Traditional CRM is a collection of silos. Those are now broken. Sales, marketing, and support use different tools and don’t talk to each other. This leads to delays and frustration.” CRM workflows that use AI to automatically resolve bottlenecks should provide the solution.
An example came from SaaS company Pure Storage, which stated that the response time to customer inquiries became seven times faster thanks to automated follow-up. “We saw an immediate difference, especially with repeated questions. The agent knows what was discussed earlier.”
Lightheartedness and leadership
Not everything was about AI, demos, or announcements. In one of the event’s most popular sessions, Jessica Williams and Conan O’Brien talked about authenticity, daring to fail, and making mistakes. Williams: “Create safe environments where people can be themselves. Don’t hire people only to force them to hide their strengths.”
O’Brien added: “Surround yourself with people who challenge you. Not just yes-men. That’s where the best things happen.”
The audience laughed, but the message resonated throughout the room. Everyone constantly talks about artificial intelligence, but it’s still people who design, monitor, and optimize the systems. “Human intervention remains necessary, AI will never automate everything.”
Ready for the start
ServiceNow’s message was clear: AI is no longer a future plan, but must be the backbone of how every work process happens. “Companies don’t want an experiment,” said Zavery. “They want something that works today and is already prepared for tomorrow.” This tone echoed in every session: AI agents are not a gimmick or hype, but digital colleagues with a substantial task list. Not futuristic, but deployable today. The foundations are there. Now it’s up to companies to get to work with it.