I've been in IT long enough to have seen more than my share of revolutions. First, IT was the builder—writing everything, running everything on premises, owning the stack from infrastructure to application.
Then came software as a service (SaaS), and IT became the broker—governing access, managing integrations, facilitating the sprawl of cloud services that exploded across the enterprise. Both eras required different skills, tools, and mindsets.
Now we're entering a third era. And the question I keep getting from IT leaders is a genuinely interesting one: With AI reshaping how work gets done across the enterprise, what comes next? The answer is bigger than most people expect.
The role of IT is shifting again, and it's heading toward a hybrid. Because AI lowers the cost of building, we'll see IT teams constructing more bespoke services where it makes sense while still brokering the best of what exists externally. But here's what's really changing: The distinctions between services and operations are collapsing.
When an employee has a problem today, it's rarely just a service desk or operations issue. It's usually both. A virtual private network (VPN) that won't connect touches the service desk, device management, network infrastructure, and possibly a configuration in your identity layer all at once.
The idea that you can cleanly separate IT service management from IT operations management was always a bit of an organizational fiction, and AI is exposing it.
When I think about the future of IT, I think about Autonomous Service Operations—a unified discipline where IT service delivery and operations run as one.
What "management" means in that future shifts too. With the explosion of shadow AI, shadow data, and the sheer velocity of change enterprises are navigating, IT's core management function becomes one of risk and compliance governance—not as a blocker, but as an enabler of safe, fast innovation.
Here's something I discuss openly with customers: The Level 1 (L1) service desk is being fundamentally reshaped. That’s not because the work is going away. The ticket backlog in most enterprises I visit is staggering, and mean time to resolve is still hours or days when it should be minutes.
Rather, it’s because agentic AI can now handle that L1 resolution autonomously. The repeatable reasoning and knowledge-driven cases—such as password resets, VPN issues, device provisioning, and access requests—don't require a human in the loop anymore.
What this means in practice is that your human L1 agents become Level 2 (L2) specialists. They stop executing the same runbooks repeatedly and start doing things that are far more valuable: improving processes, writing better playbooks, building automations, and pushing those down to the AI agents handling execution. That's a productivity unlock that compounds over time and makes the job more meaningful.
On the operations side, the same convergence is happening. The fragmented L2 silos—desktop, application, infrastructure—are collapsing the same way cloud collapsed compute, storage, and networking into one. Your AI-driven operations team works across all of it, identifying automation opportunities and acting on them fast.
The market is full of point AI solutions—purpose-built tools promising to solve specific IT problems with AI at the core. Some of them boast impressive capabilities. They can be tempting, especially when you can stand something up quickly and show early results.
Also, there are general-purpose AI platforms that can theoretically execute any use case in the enterprise, including in IT. In its new builder avatar, IT can get excited about the art of the possible.
When AI acts autonomously—resolving incidents, making changes, provisioning resources—someone has to be accountable for what it did, why it did it, and whether it did it correctly.
Point solutions optimized for speed often treat compliance, audit trails, and access controls as afterthoughts. General-purpose AI platforms do not have the full IT context to deliver these controls, so they have to be built on top, which no one likes to do.
That's a big problem. Regulators don't accept "the AI decided" as an explanation. Your security team can't remediate a breach they can't trace. And when something goes wrong at 2 a.m., and it will, you need to know exactly what changed, what triggered it, and how to roll it back.
Without deterministic guardrails, AI is a governance liability at enterprise scale. With them, it’s a superpower. It’s AI with a backbone: generative and agentic AI that reasons, recommends, and acts, but always within defined workflows that enforce the rules your enterprise runs on.
Every AI action is logged, every change is attributed, and every escalation follows a process your compliance team signed off on. AI moves fast; the platform ensures it moves safely.
Data presents another major challenge. Any serious IT service or operations capability runs on a rich, accurate, connected model of your environment: services, assets, dependencies, and history.
Point solutions bring their own data model, which means fragmented context scattered across tools that don't talk to each other. You get AI that's smart in isolation but blind to the broader picture: confidently wrong recommendations that solve one problem while creating three others downstream.
General-purpose AI platforms often give rise to point solutions built by various teams, which become hard to maintain and connect as the sprawl happens. These bespoke applications are good for generating personal dashboards or taking isolated actions, but they’re difficult to scale in the enterprise.
A unified platform built for IT use cases on a shared data foundation means the AI reasoning about an incident has full context: the affected resource, its dependencies, recent changes, historical patterns, open problems, and the compliance requirements governing how the fix gets applied. That's not just better AI; that's AI you can trust.
I'm genuinely excited about what we're bringing to market. ServiceNow has spent years building the data foundation: Configuration Management Database (CMDB), now evolving into Service Graph. It gives enterprises a single, AI-ready model of their environment leading to a holistic Context Engine. It's the substrate that makes everything else trustworthy and connected.
On top of that foundation, we're delivering a truly AI-native service operations experience:
- A conversational requester experience that meets employees and developers where they are—in Teams, Slack, EmployeeWorks, voice, or any channel of their choice—connected to world knowledge and your internal knowledge bases, resolving issues before they become tickets
- An L1 IT Service Desk AI Specialist that autonomously handles incidents and requests without human intervention and escalates genuinely complex cases to your human L2 specialists, all within governed workflows that satisfy your audit and compliance requirements
- A unified Autonomous Service Operations layer that brings together IT Service Management and IT Operations Management so that your human specialists can focus on improving systems rather than executing repetitive tasks across disconnected tools
- Proactive operations intelligence, which correlates alerts, metrics, and events across your monitoring ecosystem and can detect early warning signs, reduce alert noise, and prevent outages before they reach your users
- Learning-enhanced automation (LEAP), which identifies automation opportunities, helps your team build them quickly with human-in-the-loop guidance, and deploys them back to the AI agents handling frontline resolution
- A workflow creation experience underpinned by enterprise vibe coding, letting your builders innovate quickly on the platform using the AI coding platform of their choice and deliver value to your entire organization with governance and predictability
- An AI-driven implementation experience that gets you to value in days, not months, already knows industry best practices, and can ingest your policies and processes to configure itself around how you operate
The IT of the future is a service operations function that runs autonomously for the routine, intelligently for the complex, and compliantly for everything in between. Getting there means choosing AI that doesn't just move fast; it moves safely, with full visibility, inside the guardrails that enterprise operations demand.
Point solutions can deliver quick wins. An AI platform delivers compounding returns. The difference shows up not in the demo, but in the audit, the incident postmortem, and the board conversation about AI risk.
We've been on this journey with our customers for years. The data foundation is there. The workflow intelligence is there. Now the AI experience is ready, and it's built to run the enterprise, not just impress in a proof of concept.
The best incident is one that never happens. The best service desk is one that runs itself—safely, transparently, and at scale.
Find out how ServiceNow can help you unify IT service and operations on one AI platform.
