You Don't Have an AI Problem. You Have a Job Description Problem.

I spent years as the CIO of a print company. I know exactly how that job goes, because I lived the version nobody puts in the org chart. You do not spend your days setting strategy. You spend them keeping the lights on. Password resets. A down press that needs its integration babied back to life. An EDI feed that broke overnight and a customer who needs their file by 10. The strategic project, the one that would actually move the business, gets pushed to next quarter. Every quarter. For years.

That is not a personal failure. It is the structural reality of IT in print and packaging. Most shops spend the overwhelming majority of their technology budget and their people just running the business, keeping legacy systems alive and answering the same requests over and over. Whatever is left for building, for growth, for anything new, is a rounding error. We have accepted this for so long that we stopped calling it a problem. We just called it IT.

This is the thing AI actually changes. Not the quoting. Not the scheduling. Those matter, but they are the surface. The real shift is that for the first time, AI can absorb enough of the reactive work that the people running your technology can lift their heads. They can stop being a help desk and start being what this industry has never let them be. Strategic thinkers. The part of the business that sees where technology can change the game, and has a seat at the table when those decisions get made.

The Metric We Have Been Using Is the Problem

For decades, we measured IT the way we measure a call center. How fast you closed the ticket. How few complaints you generated. How much uptime. Those are survival metrics. They tell you the plant did not burn down. They tell you nothing about whether technology made you any money.

Here is the metric I want this industry to adopt instead: time returned to the business. When AI takes the reactive load off your team, the only question that matters is where those hours go. Not back into the ticket queue. Not scattered across busywork. Into the work that was always getting shelved.

In print, time returned is not an abstraction. It is press hours you stop losing to last-minute rescheduling. It is estimating that takes two hours instead of two days. It is a CSR who spends the afternoon selling instead of rekeying the same job into three systems. It is the estimator's judgment captured before he retires. Those are P&L numbers, and your plant manager already watches them. If AI does not move one of them, it is just entertainment. I have said that for years and I am not softening it now.

Stop Buying Technology the Way We Always Have

Here is where our industry has to look in the mirror. Print does not run experiments. Print runs multiyear system replacements. We form a committee. We write a requirements document the length of a novel. We sit through demo after demo. We sign for a platform that will not go live for a year and a half, and by the time it does, the requirements have changed and the technology we scoped it against is two generations old.

Most technology bought that way ends in regret. Ask anyone who has lived through an ERP or MIS replacement whether it delivered what the sales deck promised. In the AI era, that approach is not just slow, it is dangerous, because you are making a decision today on data that will be a year stale before anyone logs in.

The operations pulling ahead are doing the opposite. Small, reversible tests. Ninety days. A clear number that defines success. A low cost of walking away. A pilot that does not work is not a failure to defend in a budget meeting. It is data that told you the truth cheaply. Design for reversibility, not permanence. That one change in how we buy would save this industry more wasted capital than any tool ever will.

You Cannot Build on a Foundation You Have Not Fixed

Now the uncomfortable part. AI is only as good as what you point it at. Clean information produces intelligence. Messy information produces your own mistakes handed back to you faster, and with more confidence.

In print, our foundation is rarely clean. The estimating logic lives in a spreadsheet that has grown legs over 15 years. The job history sits in an MIS nobody has fully trusted since the last upgrade. The pricing rules live in one person's head. That is not a reason to wait on AI. It is the first work AI readiness actually requires. Before you accelerate, get the foundation right. Every dollar you spend fixing it comes back the first time a model runs against data it can finally trust.

I Do Not Want Your IT Team Building AI. I Want Them in the Room.

Here is where I break from the popular version of this story. The trend right now says AI frees IT to become builders, to write their own tools, and stand up their own systems. For a software company with a bench of engineers, fine. For print and packaging, that is the wrong ambition, and chasing it will waste the exact capacity you just freed.

I do not want your IT team building AI solutions. Building is not their highest use, and honestly, it is not their job. The AI that actually moves your business should be built by people who do this for a living, with the governance and accountability that comes with it. What I want from your IT team is the thing they have never had the time or the standing to do. I want them finding the opportunities.

Think about who in your building actually knows where the money leaks. Where the same job gets rekeyed three times. Where an estimate walks out the door with the wrong margin. Where a press sits idle because two systems do not talk. Your IT leader sees all of it, and for years they have been too buried in tickets to say a word about it in a meeting that mattered. Give them their hours back and point them at the business, and they become your best source of where AI should go first. Not because they can code it. Because they can see it.