How to drive value with AI: On deploying AI that actually works, with Ammex's CIO Chad Ghosn

20 years, five industries, and a career spanning Dell, Expedia, Kaiser Permanente, and the US Air Force—Chad Ghosn knows that IT’s power lies in its impact, not its complexity. He has built just about every facet of IT infrastructure we can imagine– network & system engineering teams, telecom infrastructure, and IT service management, of course.

But his defining belief? True innovation doesn’t live in IT alone but flows across departments to dissolve cross-functional silos and drive enterprise-wide growth. Now, as Global CTO and CIO of Ammex Corporation, he's putting this philosophy into action by using AI as a business enabler, and not as a back-office IT tool.

This episode is special because, for the first time ever, we have an Atomicwork customer on the podcast, talking about why they saw Atomicwork as the solution to their IT challenges. Chad speaks to our Chief Digital and Business Officer, Lenin Gali, on leading an enterprise in using AI to stay ahead of the curve.

You can listen to the full conversation here.

On leading the AI charge at Ammex

When Chad’s team started exploring AI last year, they brought in experts from across industries and academia to reshape their internal conversations, align on a bold AI vision, and take a hard look at where they stood in their digital transformation journey.

What we learned is pretty much the reality check every company will eventually face during their AI transition: the data isn’t as clean as we thought. Turns out, there’s a lot of dirty data lingering in old policies and documents that haven’t been updated in ages, which left us with a bunch of outdated processes and systems. You can’t rush good data integration.  - Chad Ghosn, CIO of Ammex

AI agents thrive on quality and scalable data, and Chad’s team understood the importance of that from day one. They’ve made sure that the information flowing into the system is nothing short of excellent.

As Chad remarks, "Our AI agent is not going to pull answers out of thin air—it needs accurate data to give you the best possible solution. So, we had to hit the reset button, rebuild all of it from scratch, and make sure the data we’re feeding into systems is solid. It took longer than expected, but it’s what we needed to do. Once we had that in place, we focused on aligning this data with company-wide workflows."

How to choose the right AI partner?

Vendors love slapping the "AI" label on anything that even remotely resembles automation, but he knows it can open a pandora’s box of complications—implementation headaches, team resistance, and potential integration nightmares. But Chad’s team was ready with a strategy before considering a tool.

We started by taking a hard look at the technology we already had in place, assessing both our current needs and what we anticipated for the future. The goal was to identify the gaps and see how we could better meet those requirements with AI.

In their surveys, service management clearly emerged as an area where AI-driven agility was crucial. Chad realized the problem: the old system was stuck in a loop where every request had to be manually routed, with people stuck in portals and rigid fields. That’s when it hit him—they needed a partner who could break free from legacy systems, move fast, and adapt to the latest tech trends. Enter Atomicwork.

In Chad's words: "Atomicwork had the right leadership, and vision, and was cost-effective without sacrificing flexibility. Now with Atomicwork, our requests get handled instantaneously, approvals are granted within minutes, and users no longer have to log into portals to check on progress."

The impact they saw was immediate and kept building momentum. Now teams could self-serve common issues like password resets while more complex queries got routed to human agents with perfect context, intent, and the right set of keywords.

Atomicwork learns, adapts, and evolves quicker than any human, fine-tuning itself with every new bit of feedback. It was a simple, impactful rollout that turned heads in the boardroom. We're aiming for an 80% deflection rate, already seeing 65%, soon to hit 67%.

The system's self-updating nature meant the knowledge base grew automatically to revert user queries– so no more manual updates, no more fragmented data—just increasingly precise, efficient service boosting productivity and satisfaction across the organization.

Read more: How Atomicwork transformed Ammex Corp's service management with AI

CIO challenges in the post-AI age

Chad admits that trusting AI with sensitive enterprise (personal and confidential) data is a tough sell, especially when the engine wasn’t built in-house. To address this trust deficit, he sat down with the leadership team to understand the current security model, recalibrated encryption and security protocols to protect employee and user data and then rolled out the AI system. "Trust isn’t static; it’s earned, reviewed, and re-earned with every interaction", remarks Chad.

Despite the hype of AI-powered workflows, the age-old challenges of scaling tech aren’t going away. At Ammex, a PPE company that grew organically through word-of-mouth for three decades, Chad faced a familiar dilemma:

As Chad reflects on his CIO journey, he also highlights the art of scaling between large and small organizations. This means carefully choosing and implementing solutions—whether CRM or cloud services—that drive revenue growth while keeping the team nimble. As Chad puts it, "You have to take what works in large organizations and figure out how to make it work with fewer resources and a much leaner team."

The AI rules that Chad lives by

If your organization doesn’t have a unified AI strategy, you're asking for chaos. Without cohesive planning, you’ll end up with disjointed solutions—tools that might work for one department but clash with others. Think processes that don’t integrate, short-term fixes, and a pile of software that’s more headache than help.

Fixing this fragmented workflow requires IT to become the architect of AI-driven change, map out how AI will impact the business, and align every investment with business goals. The real challenge, however, will be to cut through the clutter of endless vendor pitches to find solutions that truly fit.

AI is evolving rapidly, much like the early days of cloud services. Just as we once worried about losing control over data in the cloud, similar concerns arise with AI—can it handle sensitive data securely? Over time, as with cloud security, AI will need to prove it can protect data and ensure it remains within the organization’s control.

“Sitting on the sidelines isn’t an AI strategy,” Chad says on building an AI-ready culture. Even if your approach is “wait and see,” make sure you have a clear plan. How will you adapt once AI finds its footing? What will adoption look like for your business? And most importantly, how will you communicate that plan to your team?

Don’t let AI catch you off guard—lay the groundwork now and stay ready to act when the time comes.

There's more where that came from. Tune in to hear Chad's complete playbook for AI-driven IT. Listen to the full episode here.