Knowledge 2026 Day 2 keynote: The blueprint for agentic business
"Agentic AI has gone from a concept to a movement,” proclaimed Amit Zavery, ServiceNow president, chief product officer, and chief operating officer, kicking off the second day of Knowledge 2026.
Where Day 1 established a vision for the future, Day 2 went deeper, walking 25,000 attendees through ServiceNow's architectural blueprint for agentic business: sense, decide, act, and secure.
The promise of agentic AI is real, Zavery said, but most organisations aren't set up to capture it. The average enterprise runs hundreds of apps, each with its own tech stack, models, and policies. Everyone is optimising their own piece, but nobody is orchestrating the whole. Zavery has a name for it: the patchwork enterprise.
This phenomenon explains why businesses are keeping their AI ambitions relatively modest. Only 20% of organisations expect to use agentic AI to orchestrate multistep workflows in the next two years, according to the ServiceNow Enterprise AI Maturity Index 2026.
The organisations seeing results share a common trait: They're running on a unified platform with governance built in. "Only ServiceNow brings together data, AI, workflows, and security in one platform," Zavery said. "That makes us the AI platform for innovation."
Build anywhere, run on ServiceNow
Innovation has never moved faster. But as prototypes proliferate across teams and tools, they create a dangerous gap between what gets built and what makes it to production. ServiceNow Build Agent closes that gap, meeting developers inside ServiceNow Studio or any external environment via the ServiceNow software development kit.
Zavery also announced an expansion of the partnership between ServiceNow and Google Cloud. Google's full product stack is now connected to the ServiceNow ecosystem, unifying governance between ServiceNow AI Control Tower and the Google Gemini enterprise platform. In addition, Google's agentic data cloud is now connected to the ServiceNow AI Platform.
"We have changed the game for app development," said Zavery.
Making sense of autonomous work
Gaurav Rewari, executive vice president and general manager of data and analytics products at ServiceNow, argued that for work to be truly autonomous, it needs a complete nervous system: one that connects, controls, contextualises, and converges data into the flow of work.
ServiceNow Workflow Data Fabric connects data wherever it lives, with read and write capabilities across structured, unstructured, static, and streaming data. And the newly released ServiceNow Data Catalog adds native metadata management and governance capabilities throughout the data lifecycle.
Finally, Context Engine then unifies decades of organisational knowledge into a single layer, feeding AI context at the moment of decision and enabling what Rewari called the "embedded, always-on AI analyst."
Decide with confidence
Seeing data and understanding it are two very different things. Most organisations don’t know the distinction, which is why so many promises of autonomy have gone unfulfilled, said Nenshad Bardoliwalla, group vice president of AI products at ServiceNow.
While large language models (LLMs) are good at finding patterns, they lack insight into the context behind how your specific business operates. "The real differentiator," he said, "is whether AI has the context to make decisions you would trust it to make on your behalf in the middle of a critical workflow."
AI that acts
At ServiceNow, 90% of IT support requests are now handled autonomously. Autonomous workflows extend across employee experience, security and risk, app development, and customer relationship management (CRM)—with orders flowing from quote to fulfilment without a manual handoff.
That’s "AI agents working on the same platform with the same data and the same governance," said Kellie Romack, ServiceNow chief digital information officer. "That's huge."
Using autonomous workflows, CVS Health has delivered 2.5 million AI conversations that have saved time for patient- and customer-facing staff, Romack said. In a demo, Niki Patel, director of AI product management at ServiceNow, showed an IT service desk manager whose AI agent specialists, onboarded just the day before, were already doing work.
Provision, control, secure
Agentic AI has changed the threat landscape. For enterprises, the question is this: How does security change in response?
John Aisien, senior vice president and general manager of central product management and security and risk at ServiceNow, took the stage alongside Yevgeny Dibrov, senior vice president and general manager of Armis at ServiceNow, and Tarun Thakur, group vice president and general manager of identity security at ServiceNow, to answer that question.
For ServiceNow, the foundation is three graphs working in concert:
- Armis powers a cyber asset graph that closes the visibility gap across IT, operational technology, internet of things, and critical infrastructure.
- Veza powers an access graph built on 120 billion mapped permissions for humans, non-humans, and AI agents, ensuring every agent operates with least-privilege enforcement at the moment of action.
- ServiceNow Knowledge Graph ties them together with the business context that determines what should, and shouldn't, be happening at any given moment.
AI Control Tower sits above all of it, discovering every AI system running across the enterprise and remediating issues before they cause disruptions.
The agentic business starts now
Zavery wrapped the session by thanking the ServiceNow community for making the business run. ServiceNow Otto, 20 new AI specialists, AI Control Tower, Action Fabric, Context Engine, autonomous security—all of these become possible when technology catches up to ambition.
The agentic business starts now.
Find out more about how ServiceNow can help you embrace agentic business.