I've been in the room when an ERP project goes sideways.
Not as a consultant watching from the outside. As the CIO holding the bag. I spent 16 years in executive IT leadership for a global manufacturer, eight of those as CIO reporting directly to the CEO. I watched us scale from M to over 0M. I presented technology strategy to the board. And I've seen what happens when ERP implementations go wrong (and when they go right).
The pattern hasn't changed in decades. Somewhere between half and two-thirds of ERP projects don't meet their original budget, timeline, or benefit expectations. Outright abandonments are rarer, maybe 5 to 10%, but the cycle of overpromise and underdeliver keeps repeating.
This is not a technology problem. It's a leadership problem wearing a technology mask.
The Same Mistakes. Every Time.
Executive sponsorship is a checkbox, not a commitment. I've watched CEOs announce transformations, approve budgets, and then vanish. The CIO gets left holding a multi-year initiative with shifting priorities, political landmines, and a C-suite that suddenly has "other priorities" when decisions get hard.
Real sponsorship means showing up to steering committee meetings. Making unpopular calls on scope. Firing the VP who refuses to adopt the new process. At one manufacturer I worked with, the CEO attended every steering committee meeting for eleven months. The project came in on time. That's not a coincidence.
Projects are chronically under-resourced from day one. Executives approve the budget for software and integrators but gut the internal staffing. The assumption is that existing employees will absorb project work on top of their day jobs. They can't. When you staff a project with people who have 20% availability, you get 20% of their attention and 100% of the delays.
Scope creep is treated as inevitable rather than a leadership failure. Every ERP project starts with a defined scope. Six months in, finance wants a custom report. Sales demands a workflow exception. HR insists their process is "unique." Each request is small. Collectively, they double the timeline and triple the risk.
Saying no is a leadership function. Most leaders abdicate it.
Organizations confuse necessary customization with resistance to change. The discipline here is brutal honesty. A good implementation team challenges every customization request with one question: if we don't do this, what specifically breaks for the customer?
If the answer is internal inconvenience rather than external impact, it's not a requirement. It's a preference. I've killed dozens of "must-have" customizations with that single question.
Vendor incentives are misaligned with customer success. The software vendor gets paid on license revenue. The systems integrator gets paid on billable hours. Neither has a financial interest in a fast, clean implementation.
This isn't a secret. It's the business model. If you want different outcomes, negotiate different terms. I've structured contracts that tied integrator compensation to go-live dates and adoption metrics. It changes the conversation entirely.
Data migration is underestimated every single time. Organizations think moving data from legacy systems is a technical exercise. It's not. It's a business exercise that exposes decades of inconsistent processes, duplicate records, and decisions no one remembers making.
"We'll clean it up during migration" is a fantasy. You either clean your data before the project or you import garbage into your new system.
AI Cannot Save You From Yourself
AI has legitimate, narrow use cases in ERP implementations. Automated testing, data profiling, intelligent process mining, anomaly detection. These tools can reduce manual effort in areas that have historically been tedious and error-prone.
But here's what AI cannot do. It cannot fix the underlying organizational dysfunction that causes ERP failures.
AI cannot make your CEO show up to steering committee meetings. It cannot tell your CFO that his custom reporting requirement will add three months to the timeline. It cannot force your business units to abandon their pet processes. It cannot align vendor incentives with your outcomes. It cannot clean 20 years of bad data decisions.
These are human problems requiring human accountability. Throwing technology at them is what got us into this mess in the first place.
The current wave of AI integration into ERP platforms risks making things worse, not better. It adds complexity. It creates new points of failure. It requires skills your team doesn't have. And it gives executives another excuse to avoid the hard work.
"The AI will optimize it."
No. It won't.
What Actually Works
The companies that succeed at ERP implementation aren't doing anything revolutionary. They're doing what everyone knows they should do but most won't.
They treat executive sponsorship as an active role, not a title on an org chart. They fund the project properly, including dedicated internal resources relieved of their day jobs for the duration. They define scope early and defend it ruthlessly. They challenge every customization request on one basis: does this serve the customer or does it serve internal comfort? They negotiate contracts that tie integrator compensation to outcomes, not hours. They invest in data quality before the project starts. They staff with their best people, not whoever is available.
None of this requires AI. It requires discipline.
And discipline is precisely what most organizations lack when faced with the politics and complexity of enterprise transformation.
Nobody Gets Fired for This
ERP failures persist because the people responsible for them rarely face consequences. The CEO who championed the failed project has moved on by the time the write-off hits. The consulting partner who promised results collects their fees regardless of outcome. The vendor blames the integrator. The integrator blames the client. The CIO becomes the sacrificial lamb.
I've been that CIO. I've inherited implementations that were 18 months late and 200% over budget. I've had to walk into a board meeting and explain why a project my predecessor approved was never going to deliver what was promised.
Until accountability catches up with the decision-makers, expect more of the same. New vendors. New buzzwords. Same failures.
AI is just the latest buzzword. It won't fix your organizational dysfunction. Only you can do that.
If You're Facing an ERP Decision
I spent 16 years as a CIO, including eight years leading IT for a 0M global manufacturer. I've lived through ERP implementations from both sides: the ones that worked and the ones that became expensive lessons.
If you're about to make a seven-figure technology decision, or you're stuck in the middle of a project that's going sideways, I'll give you 45 minutes to diagnose the situation and tell you exactly what I see.
The 45-Minute IT Strategy Session:
- I'll assess your biggest technology risk or decision
- You'll get specific recommendations you can act on immediately
- You'll walk away with clarity on whether your current approach can support your next stage of growth
Even if we never work together, you'll leave with at least one actionable insight you can implement immediately. If you don't, I'll personally connect you with someone in my network who can help.
No pitch. No pressure. Just a straight conversation with someone who's been in your chair.
[Book the Strategy Session]
Raphael Savastano is the founder and principal consultant of ROFONIC LLC. With 25+ years in IT, 16 years in leadership, including 8 years as CIO scaling technology for a global manufacturer from M to 0M. He now helps growing companies get executive-level technology and operations leadership without the full-time cost. Want to know where your technology actually stands? Take the Founder's IT Reality Check →
