Introducing Certified AI Coworkers — production-ready from day one.
read more
Customers

In This Guide:

Share Article

Summarize with:
Updated:

How ITSM Evolved in 2026 and What to Expect in 2027

AI has evolved with ITSM over the years. Until recently, most AI features answered questions and suggested articles, and a person did the rest. Now AI agents can resolve requests, approve access, and update records. Gartner expects 40% of enterprise applications to include task-specific AI agents by the end of 2026, up from less than 5% in 2025.

Take a common request: an employee needs access to a tool. A virtual agent that only answers questions will often point to a knowledge article, and the employee still raises a ticket and waits for IT to approve it and grant it. An AI agent that can act, checks the request against policy, sends it to the right approver, grants the access once it's approved, and closes the ticket. IT steps in only for exceptions.

This guide covers the five shifts that changed ITSM in 2026: AI-ready data, AI that acts, AI governance, cost accountability, and enterprise service management. It also covers what teams gain from each and what to expect in 2027.

Top 5 ITSM Shifts in 2026 and Where Atomicwork Stands

1. AI Is Only as Good as Your Data

After a year of AI pilots, IT teams are finding that AI is only as good as the data and processes underneath it. A CIO consultant, Bryce Austin, says that AI can perform well when it has enough information, but can be a waste of time for IT teams without the data and AI readiness to use it.

So what are the fundamentals? At Atomicwork's FUSION'26 conference, Chris Asign, VP of Business Technology, Data Engineering and AI Transformations at Ironclad, listed them: "I kept coming back to the basics. Clear ownership. Trusted data. Strong identity and access controls. Workflows that reflect how people actually work. Governance that is built into the operating model."

This is one of the top reasons customers choose Atomicwork. Atomicwork supports requests, incidents, changes, assets and configuration in one platform. Universal Context adds the history of your users, assets and past support, so AI Coworkers work from the same information your team does. And because every AI Coworker is governed for what it does, with a budget, a manager and an audit trail, you stay in control.

2. AI that Acts, not just Assists

The second trend is the move from AI that suggests to AI that finishes the job. ISG expects that through 2027, 50 percent of enterprises will adopt ITSM software with agentic AI for proactive issue detection, incident resolution, and workflow orchestration. In practice, that means AI that spots a failing service, restarts it, confirms the fix, and closes the incident, with a person only for approvals and exceptions.

An AI Coworker does this kind of work. Say a service starts failing overnight. The Coworker spots it, restarts the service, confirms the fix, and closes the incident, and it asks for approval before any change that needs one. It works the way a new hire does: it has a name, a defined scope, the tools and access it needs, a budget, and a human manager. The newly launched certified AI Coworkers in the Atomicwork AI Coworker Marketplace are pre-built for IT and security work. They're self-improving over time, and each one is tested against a role-specific benchmark so it gets better at the job it does.

3. AI Governance

AI used to answer questions. Now it can reset accounts, approve access, and update records. That makes it important to decide what AI is allowed to do, and to keep a record of what it did.

As AI adoption matures, AI governance holds strong as a core priority to manage risk, regulatory audits, and responsible AI guardrails. Forrester predicts 60% of Fortune 100 companies will appoint a head of AI governance. Only 16% of IT professionals say their organization fully trusts AI to make and execute operational decisions. That caution makes sense once AI can approve access or change a record.

Atomicwork treats every AI Coworker like a member of your team. Each AI Coworker has a defined role, managed credentials, bounded access, a compute budget, and full audit trails. Atomicwork is SOC 2 Type II certified and holds an ISO 42001 certification for AI management systems.

4. Value, ROI, and Cost Accountability

If AI is doing real work, someone should own it, cap its cost, and show what it delivers. As AI has made the work easier for IT teams, the returns on investments have started to show. In Atomicwork's State of AI in IT 2026 survey, 67% of IT professionals reported positive ROI and only 3% reported negative ROI. A year earlier, only 24% reported positive ROI.

Finance leaders are now closely involved in reviewing AI costs. In Deloitte's Q3 2026 survey of 200 CFOs at North American companies with at least $1 billion in revenue, nearly 80% said finance has a formal role in AI-related decisions. Keeping costs under control once AI is running is where many teams fall short: about 53% review AI costs only after the fact, and only 7% can pause overrunning usage automatically or in near real time.

We think the way to manage AI costs is to count outcomes such as resolved requests, and Atomicwork prices on that basis. AI Coworkers are priced based on use or outcomes achieved by them; they are included in all pricing plans, with usage-based and outcome-based pricing available. With outcome-based pricing, you pay per resolved ticket, such as an access request handled end to end, so the cost of each outcome is set in advance. Each AI Coworker also has a budget and a human manager, and built-in analytics let teams track how they perform.

5. Enterprise Service Management (ESM)

Enterprise service management (ESM) takes the way IT handles requests and applies it to HR, finance, and facilities. For employees, it means one place to ask, whichever team ends up doing the work.

It's not common yet. 74% of organizations use AI in at least one service management team, but only one in five has fully embedded it across all of them, from IT and HR to Finance.

Atomicwork was built to work across those teams from the start: enterprise IT, HR, Finance, and Workplace, with a Universal AI Coworker for employees. An HR Ops Specialist, for example, helps employees with onboarding, policy questions, benefits, and time off.

What are The Benefits of Adopting These Strategies?

Each shift brings its own benefit. Getting the fundamentals right means AI works from accurate records and clear processes, so its answers and actions are more reliable.

AI that acts takes repeated requests off your team, so people spend their time on harder problems. AI governance means you know what AI can access and what it did, so you can give it more work with confidence. Cost accountability puts an owner and a limit on AI spending, and gives you results to show finance. Enterprise service management gives employees one place to ask for help, whichever team does the work.

A customer example shows what this looks like day to day. Pepper Money, an Australian financial services company, went live with Atomicwork about four weeks after signing. Resolution time dropped 75%, from 29 hours to about 7. Nearly 40% of requests are now fully automated end to end, and onboarding that took five business days now takes two hours.

What to Expect in 2027 ITSM Trends?

  1. AI will work in teams. By 2027, one-third of agentic AI implementations will combine agents with different skills.
  2. Zero-touch support will spread, starting with routine requests. Gartner expects that by 2030, 50% of organizations will use AI agents with minimal human involvement in IT operations, up from nearly 0% in 2026.
  3. Experience-level agreements (XLAs) will be given importance. Traditional SLAs will remain important, but IT teams will increasingly complement them with measures of employee experience and business outcomes.
  4. AI-caused incidents will be a category of their own. We expect IT teams to log, investigate, and learn from incidents caused by an AI's actions, not only by people or systems.

Ready to see how Atomicwork can ease your IT Service management?

See how Atomicwork's AI Coworkers handle real service desk work. Talk to our expert

Heading

This is some text inside of a div block.
Meet 100+
tech-forward CIOs
Date icon for Atomicwork event
Sept 24, 2025
Venue icon for Atomicwork event
Palace Hotel, SF
Request an invite

Frequently asked questions

Will AI take over 100% of IT service desk work?
Has AI governance improved?
What is the difference between generative AI and agentic AI in ITSM?
Which IT tasks should you automate with AI first?
Do you need to replace your ITSM tool to use AI agents?

More resources on modern ITSM

No items found.
No items found.
Text Link
This is some text inside of a div block.