In this installment of the IT Leaders Fast-5 — InformationWeek’s column for IT professionals to gain peer insights — Ed Fox, CTO of MetTel, recounts why adding AI on top of a broken process only makes things worse, faster. “If your process is broken to begin with, then [AI] is not going to help you,” Fox said.
At MetTel, most AI remains focused on automation. As Fox pushes his teams to experiment more broadly with AI, he is also tightening oversight after shadow AI started showing up across the network. He is now formalizing that push through biweekly AI productivity sessions and quarterly reviews designed to raise the organization’s overall AI fluency.
Fox has worked at MetTel for more than 25 years and is a member of the Forbes Technology Council.
This column has been edited for clarity and space.
The Decision That Mattered
What recent decision — technical or organizational — has made the biggest difference, and why?
I recently made a decision to hire a senior-level employee who had AI experience when I could have used his salary to get three working analyst engineers. That was a big deal.
We’ve been on the automation journey and the machine learning journey since 2016, and this [hire] gets us to the next level.
For the most part, we’ve used machine learning and AI to look at our unformatted data and give us responses. Once we’ve looked at that, [we realized that] 98% of [AI] is automation. That’s primarily on the network operations side, not customer-facing or customer service. It’s used to open up [service] tickets.
The AI piece was learning what we could automate. In our business — where we have to be very cognizant of opening tickets — AI is essentially automation. Using productivity apps to draft an email quicker is a different type of AI, but for us, it’s mostly automation.
The Hard-Won Lesson
What didn’t go as planned recently — and what did it force you to rethink?
We tried to deploy AI — we put an automation process together and quickly realized, in this one instance, that all we were doing was making our mistakes quicker with customers. We tried to automate or put AI into a manual process that was broken, and we thought AI and automation would fix it. But it just made it worse, quicker. If your process is broken to begin with, then [AI] is not going to help you.
We went back, put certain people on the project and the process, and had them take lots of notes with the help of an AI reader. In that sense, AI helped, but we fixed the process manually.
It would have saved us hundreds, maybe thousands of hours if the process, workflow, pivot points and decision-making points were defined first.
The Talent Trade-Off
Where are you investing in talent right now — and what are you consciously not investing in?
So, I made that decision to hire that [AI executive]. But I’m looking to bring up the average [AI talent] of everyone involved in the organization. Recently, I put this [proposal] out to my team — that every other Thursday I want to have an AI meeting where people come and show us what they’re doing personally [with AI], regarding productivity for MetTel or for our customers. The every two week demo has been in place since last year.
The teams involved will include my group, our global NOC, customer service, network engineering, the network backbone team, mobile core team, mobile core security, compliance team — it’s across the entire organization.
I’m also putting together a quarterly meeting — the first one starts next month — with my six direct reports and their team, where a different department each quarter shows us what they did with automation.
We also need more product managers and more people that are more focused on customer service.
The External Signal
What recent external development is most likely to change how your organization operates, even indirectly?
The memory [RAM] issue is definitely impacting us. We see anywhere from a 10% to 30% increase in [cost of] the hardware that we’re selling to our customers. A lot of those are three-year contracts so that’s difficult for us to navigate through.
Regarding AI — OpenClaw (formerly Moltbot, Clawdbot) is very scary — it’s so agentic. I’m proactively telling everyone to get involved in [AI] but all of a sudden I see it on my network and think “Oh my god, what did I do?” We’ve teamed up with Netskope that [monitors shadow AI] and tells you what AI everyone is using and what’s in that AI — if it’s any internal information. But that’s automation, too, so it’s about trying to keep up.
The Perspective Shift
What have you read, watched, or listened to recently that changed how you think about leadership or technology — even slightly?
We had futurist Zach Katz speak at one of our client advisory boards and at our Innovation Summit, reading his book “The Next Renaissance: AI and the Expansion of Human Potential.” He’s pretty realistic about what AI can do, and he also attaches it to social issues. I like reading that kind of book to figure out where we are now and what’s going to happen next.
