The CIO’s next job: Eliminate the work, then run the agentic AI workforce
SaaS gave the business speed and left IT with the complexity. AI is IT’s chance to change that bargain.
About 15 years ago, SaaS made the enterprise a deal. Every function could adopt new technology and move faster without waiting on IT. The business gained speed, and IT inherited the complexity.
Look at what that deal produced: thousands of applications and integrations, point solutions wedged into the gaps between systems of record, and a huge bill just to corral and govern all the data they generate. It’s a lot to carry, and IT has been carrying it quietly for a decade and a half.
Now AI is starting another shift, and it’s bigger. We’re moving from “let me answer that question for you” to “let me just go get that done for you.” That’s why I think this is one of the most exciting moments to be a CIO in decades. But IT can’t lead this transformation while it’s still buried in yesterday’s work.
We’ve been measuring the wrong thing
For decades, ITSM has gotten very good at routing tickets, prioritizing them, escalating them, and timing how quickly humans resolve them. The ticket became the unit of success, but it’s just the container we use to organize work. Around it sits an enormous amount of routine, repetitive effort.
AI lets us ask a different question. Not “how do we close tickets faster?” but “how much of this work can we remove from the system entirely?”
That’s a fundamentally different ambition from making ticket management 20% more efficient. It means measuring work eliminated, not tickets automated.
Watch on demand: What's new in Freshservice
What eliminating work looks like
At Freshworks, we’ve boiled our product strategy down to one question: for every role across the IT ServiceOps lifecycle, what are the most common human interventions, and how do we dramatically reduce the human effort involved? That’s the whole strategy. And how do we get there? Three investments are moving us toward it:
Context. AI is smart, but it doesn’t inherently know which laptop an employee has, what they’re entitled to, or which infrastructure supports a critical service. That’s why we’re investing in Freshservice ITAM and the deeper asset intelligence from Device42. The more AI knows about the environment, the more reliable it gets.
Ease. Process mining helps teams see how work actually gets done and find automation opportunities they’d otherwise miss. In early access, Freddy AI Agent Studio lets teams describe a workflow in natural language instead of building every step by hand. No specialized AI expertise required.
Action. We’ve expanded Freddy AI Agents into email, and Freddy AI Copilot is becoming more proactive, recommending and initiating supported next-best actions with human oversight.
I don’t want to hand IT a better chatbot. I want AI that understands the job to be done and helps finish it.
How McLaren, New Balance & more turn service performance into a competitive edge
What you do with the time you get back
The harder question is what happens to the capacity you free up. Our own CIO, Ashwin Ballal, has been cutting the repetitive work his IT organization performs. Rather than treat that capacity as headcount to trim, he built a team of forward-deployed IT engineers who work directly with business functions to help them build their own AI agents.
Think about the shift. IT stops spending its days responding to requests and starts helping the business change how work gets done. That’s the first chapter of the CIO’s transformation: Fix IT’s own work first.
The second chapter: A workforce IT didn’t hire
The second chapter is bigger. Someone in finance can already build an AI agent. So can someone in HR. Employees are bringing increasingly powerful personal AI tools to work, and soon every one of them will walk in with an AI in their pocket. IT can’t, and shouldn’t, try to stop that innovation.
But picture a company with hundreds, eventually thousands, of agents doing work. Who knows what they’re doing, what they can access? Whether they’re following policy, delivering value, or what they cost to run? That isn’t just an AI problem. It’s an IT management problem.
It’s also an economic one. Labor has historically been one of a company’s largest but most predictable expenses. Now imagine AI agents executing millions of jobs around the clock, consuming inference, compute, and data with every action. Earlier this year, I wrote about how that turns software spend into something closer to a metered labor cost, one that finance teams will have to forecast and measure in ways they never have for traditional SaaS. Managing it will become one of the CIO’s defining responsibilities.
That’s why Freshworks recently announced its agreement to acquire Torii. A need to understand applications, AI consumption, and software economics only will grow. It’s also why I see Freshservice as bigger than service management alone. We want to help IT transform its own work, and over time, help IT lead the transformation of work across the enterprise.
Read more: AI is changing the economics of software. IT needs a new playbook.
A question to take back to your team
If you gave a meaningful portion of your IT team’s capacity back to them, what could they build? And when AI agents arrive, which function will make sure they’re visible, governed, and worth what they cost?
For CIOs, the answer to both questions can be the same: IT.
