On a recent episode of AI Impact, host Russ Fradin sat down with Mike Giresi, CIO of Vertiv. Vertiv builds the power and cooling infrastructure behind the world's AI data centers, and their hyperscaler customers are spending hundreds of billions a year on AI infrastructure.
This raises an obvious question: what does the CIO of that company think about running AI inside his own 30,000-person organization, with operations in 100 countries? Mike's answers are strikingly practical, and they point out where most enterprise AI programs quietly stall.
AI proficiency versus AI theater
Mike has spent real time on a problem every CIO and CTO is grappling with: how do you tell the difference between AI that's producing slide decks and AI that's producing value? He uses a two-part test:
- Are we doing something materially faster? Is work that used to take 25 days down to 15 minutes?
- Does the increased speed translate into value the customer or the employee actually feels?
Mike was deliberate that not every win shows up in the P&L. Some of Vertiv's strongest use cases save time rather than headcount, and Mike treats this as legitimate—but only because the gains are visible and measurable. "We are not trying to look at AI as, 'Let's just put it against everything and eliminate half the company'," he said. For Vertiv, the goal is speed and accuracy, not staffing cuts.
The trailblazer problem and the CHRO partnership
One of Mike's sharpest observations is that most large enterprises have already proven AI works inside their walls. Engineering at Vertiv was way out ahead of the rest of the business on adoption, exactly because the customer base demanded it. The problem isn't the absence of trailblazers; it's the gap between the 10 percent who have figured it out and the other 90 percent who haven't.
That's where Mike's view of the CIO-CHRO relationship has shifted. He doesn't see these roles as adjacent anymore. Like others on the cutting edge, Mike is building a digital workforce, with each human partnering with one or more agents. But this requires the CIO and CHRO to jointly define and approve roles, skills, incentives, and org structure.
Mike candidly admits that today's agents are still mostly vertical, and that horizontal agent-to-agent integration across functions like integrated business planning isn't here yet. He won't even promise that it will be here in six months. But the org design work, he argued, can't wait for the tech to catch up.
Governance as education, not lockdown
Asked about cybersecurity and governance in an AI world, Mike pushed against the instinct to clamp down. Vertiv's approach starts with education: a strong, proactive communication program around cyber risk, with monitoring and incentives rather than punishment.
On vendor governance, anyone in the company can bring an idea forward, but tools route through an architecture review run by the operating team: operations, manufacturing, engineering, and tech, with HR and legal pulled in when needed. What Vertiv explicitly avoids, Mike said, is a steering committee of people disconnected from the business "evangelizing" tools the operators don't understand and can't use.
Building the partnership with the CFO
For CIOs trying to fund AI work, Mike's advice on the CFO relationship was disarmingly simple. Help the CFO succeed at what he or she is personally accountable for first. Once a CFO sees the value AI brings to finance itself, they extrapolate it across the company on their own. Mike also flagged the importance of an independent value-realization function; that is, people tracking AI ROI who aren't the same people implementing the changes. Controllers respond to data, not enthusiasm, and giving them clean, unbiased data buys CIOs the freedom to do more.
What to tell CIOs who feel behind
Mike was the first employee at an online ad network in 1996, before Netscape went public. He doesn't think this moment compares. The dot-com boom was largely about brand sites and ad-dollar shift. This wave, he argued, touches everything; there's almost nothing in the enterprise stack that isn't potentially disruptable.
Russ noted that, at a recent conference, 80 percent of CIOs said they felt behind on AI.
Mike's prescription for this anxiety is practical. It starts with asking the Simon Sinek question: "Why?"
If the gap is data capability, accelerate it. If it's culture, show people: build an MVP, take leadership through it, let them touch the value.
Quoting Coinbase CEO Brian Armstrong: "Action produces information." Most failed tech investments, Mike said, fail on people, process, and culture, not technology. Enterprise CIOs and CTOs who internalize that, and pair governance with education rather than lockdown, are the ones who'll come out of this wave ahead.
Listen to the full conversation between Russ Fradin and Mike Giresi on AI Impact.