ServiceNow has launched AI Experience, a multimodal interface that puts AI agents, data and workflows behind a single entry point on the Now Platform. The company pairs agent optionality with strong governance and extends the approach into an autonomous CRM that aims to speed resolutions and quoting while improving customer and employee experiences.
Enterprises have accumulated numerous tools that rarely integrate seamlessly. Users find themselves juggling multiple logins and tabs, while critical information remains isolated in separate systems.
According to ServiceNow’s leadership, AI will only address these issues if it is built directly into the platform and focused on completing tasks. The AI Experience is envisioned as a new user interface for work, where prompts replace traditional pages and conversations substitute clicks. The goal is a unified interface that understands context, acts accordingly, and keeps users in control.
A single front door for work
“AI Experience from ServiceNow is addressing one of the biggest challenges enterprises face today, fragmented and clunky user experiences that slow down work,” said Amy Lokey, Executive Vice President and Chief Experience Officer. “By creating a unified, contextual and intuitive AI Experience for the enterprise, we are putting AI into the flow of work, meeting users where they are and empowering them with access to workflows, data and AI agents.”
Amit Zavery, President, Chief Product Officer, President and Chief Operating Officer, framed the shift as structural. “We are at a critical crossroads in technology. AI does not just recommend, it acts,” he said. “This is not a layer bolted onto legacy tools. We have built an AI native platform on one unified data model, so outcomes are deterministic even when agents work in flexible ways.”
Agents that act across voice, web, data and screen
AI Experience surfaces five capabilities in one conversational canvas: AI Voice Agents, AI Web Agents, Build Agents, AI Data Explorer and AI Lens. Voice Agents handle approvals and troubleshooting with complete authentication.
Web Agents navigate third-party sites and apps like humans, clicking, filling out forms, and orchestrating flows when formal integrations are lacking. Build Agents enable teams to describe an app in natural language, then scaffold tables, flows, and interfaces within the modern ServiceNow Studio.
Data Explorer analyses operational data in context and proposes fixes without interrupting the workflow. AI Lens turns a screenshot or on-screen artefact into a plain English diagnosis or next step.
“We are moving from clicking through menus to interacting with technology that understands what a user needs and helps them get it done,” said Lokey. “Prompts replace pages, conversations replace clicks, and AI becomes a teammate.”
Deterministic meets agentic
A central theme was the blend of probabilistic power with deterministic control. “We combine agent reasoning with deterministic workflows so every AI-driven action is executed, logged, explained and auditable,” said Dorit Zilbershot, VP Product Management, AI, ServiceNow. Lokey added that users can see the reasoning steps and the team of agents assembled for a task, with humans retained in the loop where required.
Zavery noted the platform can check outcomes against insights from two decades of workflow data to confirm that agent results align with enterprise intent.
Governance and model choice
ServiceNow is treating AI like any other enterprise asset. AI Control Tower is the hub for discovery, onboarding, monitoring and governance of agents and models, whether built on ServiceNow or elsewhere. Organisations can define AI policies, evaluate agents during design using datasets, and judge models; then, sample behaviour at runtime for quality, risk, and compliance. “People stay in control and we give them optionality at every level,” said Zilbershot.
Model choice is explicit. Customers can use platform-native models, select major third-party providers, or bring their own. “Choice can be made globally or at a granular skill level so the right model can power the right task without compromise,” said Zilbershot.
Interoperability across ecosystems
Many firms will run agents from different teams and vendors. ServiceNow’s agent fabric is designed to make these agents collaborate and delegate across systems.
“Our goal is a holistic operating model for agent interoperability,” said Zilbershot. “We support approaches such as A2A and MCP and also integrate directly where standards are still forming, while preserving full traceability.”
Zavery underlined the open stance. “We are active with the community on new standards. We will not wait for every standard to settle before customers can connect agents and automate end-to-end.”
What it looks like in practice
The demonstrations focused on outcomes. A manager who needed to approve an offer while locked out of the portal used a Voice Agent that already knew the context. The agent approved on her behalf, explained there was a local VPN outage and filed an IT ticket.
Terence Chesire, Vice President for CRM and Industry Workflows, stressed that identity checks occur first. “We will never shortcut user authentication,” he said.
When an expired New Relic integration blocked access to network telemetry, a Web Agent reauthenticated, restored the link and streamed data into dashboards.
A second agent identified a recent software update as the root cause and proposed a rollback for human approval. AI Lens captured an error screenshot from a support inbox and provided an immediate explanation along with a suggested fix. With Data Explorer, a support lead requested an analysis of critical incidents, and the system surfaced patterns with recommended process changes during the same conversation.
Autonomous CRM moves from the filing cabinet to a system of action
“Legacy CRM is broken,” said Chesire “What was meant to provide a 360 view has become a digital filing cabinet. Employees put data in and get very little value out.” ServiceNow’s autonomous CRM aims to replace passive record-keeping with a system that resolves issues and accelerates growth. In service, AI agents deflect routine work and escalate what matters.
In sales, a new configure price quote capability helps reps propose the right solution quickly by matching customer needs with pricing and compatibility rules.
In a scenario for a software company operating on usage credits, agents reviewed entitlements and policies, executed a playbook to grant a one-time credit extension, and then connected the customer to sales for a tailored upgrade.
The seller viewed the service context in the same view and created a single quote that included one-off, recurring, and software items.
Zavery positioned the launch as a continuation of the firm’s workflow heritage. “We are transforming how workflows operate across people, processes and systems. We have made it effortless by bringing clean data and AI together on one platform,” he said. “The market does not need more noise. It needs results.”
Zilbershot emphasised control. “Enterprise-ready AI must be governed and secured with robust oversight to ensure compliance, trust and accountability,” she said. “With Control Tower and distributed intelligence, we enable intelligence everywhere without creating silos.”
Lokey tied the design back to user experience. “There is no jumping between tools and no new training burden,” she said. “It feels intuitive like working with a teammate who understands your style.”
Customer voices
Early adopters reinforced the value of unifying AI, data and workflows. “ServiceNow is helping transform how our technology teams serve Adobe employees by connecting AI, data and workflows across the enterprise,” said Toni Vanwinkle, Vice President of digital employee experience at Adobe.
“With ServiceNow AI Agents, we will be able to anticipate and prioritise service requests, automate resolutions across systems and deliver real-time insights so our teams can personalise support at scale.”
“As an early adopter of the ServiceNow AI Platform, we have seen how agentic AI can help us adapt quickly, respond faster and make work simpler and better for our employees,” said Brian Eble, principal in enterprise technology at EY. “With AI Experience now available, we are excited to give our people a seamless new way to interact with AI that is personalised and proactive.”
The pitch is clear. Put a single conversational interface in front of people, data and workflows. Let agents act across systems while the platform guarantees security, trust and audit. Combine model choice with control. Replace swivel chairing with a system of action. If adoption matches the ambition, the enterprise UI may finally catch up with the way people actually work.
