I was standing on the mini couch in my office.
Not because anyone asked me to. Because I had spent days mapping out the core macro processes of what a merged company would actually look like, taped them up across one wall of my office floor to ceiling, and I needed the altitude to see the whole thing at once.
We were planning a merger of equals. Two companies, two leadership teams, two technology stacks, two sets of operating assumptions. Every meeting had gone the same way for weeks. Each side argued for their tools. Each side defended their vendors. Each side justified their customizations. Each side was certain that the way they already worked was the way the combined business should keep working.
Nobody was talking about the work itself.
Standing on that couch, looking at the macro processes from floor to ceiling, I saw what the conversation had been missing. The tools were not the company. The vendors were not the company. The customizations were not the company. The macro processes were the company. They were what served the customer, what created the value, what justified everything else. And almost no one in either organization was looking at them.
That moment crystallized something I have watched happen in company after company since.
Somewhere along the way, leaders get confused about which drives which. The macro processes are supposed to be the thing the systems serve. The customer promise is supposed to be the thing the operating model enables.
But the relationship inverts. The tailored system, the chosen vendor, the implemented workflow becomes the target to defend. The macro process becomes whatever the system happens to allow. The customer promise becomes what the configuration can deliver. Companies stop running their processes through their systems and start running their systems through what their processes used to be.
That inversion is the Mirage.
The Pattern I Keep Seeing in AI
Years later, the same dynamic is playing out across every AI conversation I sit in on.
Companies are investing heavily in AI. Pilots are everywhere. Budgets are real. Boards are demanding progress. Confidence in AI ambition is high.
And measurable, durable impact on how work actually gets done is rare.
Most leaders read that gap as a technology problem. The models aren’t ready yet. The data isn’t clean enough. The vendors are overpromising. Next quarter, with the next release, the gap will close.
That reading is wrong. Or more precisely, it is incomplete in a way that lets the real problem hide.
AI is not failing because the technology is immature. AI is failing because it is being applied to the wrong layer of the business.
Leaders are aiming AI at the tools, the systems, the implementations, the same layer they have been defending and customizing for years. They are not aiming AI at the macro processes those tools were ever supposed to serve. Because in most companies, nobody can describe the macro processes clearly enough to aim anything at them.
That gap was a quiet inefficiency for decades. AI turns it into a strategic risk.
Why AI Exposes What Nothing Else Did
Most enterprise technology is forgiving of ambiguity. ERP systems hold the workarounds. CRM tools accommodate the exceptions. People fill the gaps that systems cannot.
AI does not fill gaps. AI amplifies whatever it touches. Apply it to a clear macro process and it accelerates value. Apply it to a tool that was supposed to serve a macro process but somewhere along the way became the process itself, and you accelerate confusion at machine speed.
This is the part most AI strategies miss. The technology is not the variable. The clarity of the macro process underneath is the variable.
I have watched companies invest seven and eight figures into AI initiatives that stall not because the models were wrong, but because the macro process underneath had quietly been replaced by the system that was supposed to support it.
Five teams had five versions of the same workflow because five different configurations had drifted in five different directions. Three roles owned the same decision because three different tools each thought they owned it. Nobody could agree on what “done” meant because “done” had become a status field in a particular system instead of a state of the actual work.
The AI worked. The enterprise couldn’t absorb it.
When this happens, the instinct is to blame the tool, the vendor, or the team. The real cause sits one layer down.
The macro process the company built itself around years ago is no longer visible. The systems are visible. The customizations are visible. The vendor contracts are visible. But the work itself, the actual end-to-end value flow that serves the customer, has receded into the background.
AI does not have anywhere to land in that environment. It can only attach to what the organization can describe. And what the organization can describe, in most cases, is the tooling. Not the work.
What This Means for the CIO
If you are leading technology in an enterprise right now, you are under real pressure to show AI progress. Boards want it. Operating partners want it. Peers in the C-suite want to know what you are doing. None of that pressure is going away.
But the worst move available to you is layering AI investment on top of the same system-first thinking that produced the problem in the first place.
If your team is targeting AI investment at the existing tools, the existing vendor stack, the existing configurations, you will get pilots. You will not get scale. You will spend the budget. You will not get the return. And when the post-mortem happens, the technology will get blamed for something the technology was never going to fix.
The better move is harder and slower.
Before you decide where to invest in AI, get the macro processes back on the wall. Not the org chart. Not the strategy deck. Not the system architecture diagram. The actual flows. How value moves. Where decisions live. Where the customer promise gets made and kept or broken. The architecture the tools are supposed to serve.
That work is not glamorous. It does not make for an exciting board update. But it is the difference between AI as transformation and AI as expensive theater.
The Question Worth Standing on a Couch For
If you take one thing from this piece, take this. The next time someone in your leadership team proposes an AI initiative, ask:
What macro process does this serve, and can the three of us in this room describe that process the same way without referring to a tool?
If the answer is no, and it usually is, the work is not to abandon the AI initiative. The work is to get the macro processes visible again first. Otherwise you are aiming intelligence at a system, not a business.
AI does not create the Mirage. It reveals it.
The companies that will lead the next decade are not the ones with the most advanced models. They are the ones that remember which drives which.
That clarity is the prerequisite. The technology is the amplifier.
Most leaders are still trying to amplify the wrong layer.
Trusted insights for technology leaders
Our readers are CIOs, CTOs, and senior IT executives who rely on The National CIO Review for smart, curated takes on the trends shaping the enterprise, from GenAI to cybersecurity and beyond.
Subscribe to our 4x a week newsletter to keep up with the insights that matter.
