25 Jun
McKinsey recently published a piece called “The End of ERP as We Know It?” The headline finding: AI agents have the potential to reduce ERP implementation effort by at least 50% and cut program duration in half.
Design, configuration, testing, training. The parts that used to eat six to nine months are starting to compress into weeks. If you’ve lived through a long, expensive ERP program, that gets your attention.
The expensive, grinding part of the work is finally getting cheaper.
But here’s what’s easy to miss.
When you automate the technical work, the constraint doesn’t disappear. It moves.
Hand the slowest, most manual stages of a rollout to AI, and you’re left staring at the one thing AI cannot configure for you: whether your people and your processes are actually ready to operate differently.
Once the build gets automated, change management becomes the main thing standing between you and value.
The bottleneck in a transformation was never really the software. It was the organization.
We just couldn’t see it clearly because the technical work took so long that it absorbed all the attention and most of the budget. Strip that away and the truth is exposed: the value of a new system comes from people changing how they work. And that has always been the hardest, least automated part of the job.
The technology was the visible obstacle. The organization was the real one.
Here’s where most companies will get it wrong.
When the tech gets faster, the instinct is to lean harder into the tech. Pick the fastest implementer. Let the agents configure more. Go live sooner.
The problem is that speed multiplies whatever you point it at.
Automate a process you haven’t thought through and you don’t get a good outcome faster. You get a bad one faster. And now it’s wired into your system of record.
Going twice as fast on an unclear process is not progress. It’s chaos with better tooling.
Here’s what actually works.
Clarity before configuration.
Before an agent designs your target state, you should be able to say what you actually do, why you do it that way, and what should change. Your people should understand the new way of working before the system asks them to live in it.
This isn’t a soft layer you add at the end. It’s the work that determines whether the speedup turns into value or into an expensive mess that everyone quietly works around.
The faster the build gets, the more this is where the game is won or lost.
AI is going to make the technical side of transformation faster and cheaper. That’s genuinely good news.
But it doesn’t change the equation that has always mattered most. A system is only as good as the organization running it.
The companies that win the next few years won’t be the ones who implemented fastest. They’ll be the ones who were ready.