AI in ERP: A Solution in Search of the Right Problems

Roger Pujol, CPIM, CSCP
Apr 20
2 min read

Updated: May 2

Data integration gears with the hand. — *What’s the problem we’re solving?*

ERP systems are the backbone of modern business—complex, powerful, and often clunky. So, it’s no surprise that AI has come knocking on the ERP door, promising to streamline, automate, and optimize almost everything. But as we start seeing AI embedded into dashboards, modules, and workflows, a question arises: Are we solving real problems? Or are we just excited to use AI because we can?

The Allure of "Smart ERP"

Vendors are racing to showcase their AI capabilities:

“Predictive analytics!”
“Autonomous procurement!”
“AI-driven scheduling!”
“Intelligent assistants!”

Sounds great. But when you dig in, many of these features feel... underwhelming. You get flashy dashboards that forecast the obvious. Or bots that can answer five generic questions and then escalate to a human. In many cases, the problem wasn’t that the ERP lacked intelligence—it’s that the process design was broken, the data was messy, or users were never appropriately trained. AI isn’t going to fix those things. At least, not by itself.

When AI Does Make Sense in ERP

That said, AI absolutely has a role to play. When used deliberately, it can solve problems that were previously unsolvable or wildly inefficient:

Demand forecasting based on real-time market shifts and historical data
Dynamic inventory optimization adjusting to supplier and lead time variability
Invoice matching and fraud detection in accounts payable
Predictive maintenance using IoT and machine learning in manufacturing modules
Natural language queries for business intelligence (think: “Show me open POs over $50k”)

In these areas, AI enhances decision-making and reduces manual effort—without being a gimmick.

The ERP Trap: Misplaced Innovation

The danger in ERP is building complexity that looks like innovation. AI often gets slotted into roadmaps for the wow factor, not the ROI. It becomes a checkbox, not a capability.

This leads to systems that are harder to use, not easier, and more expensive to maintain, not less. Consultants and IT teams are left holding the bag, trying to explain why the “smart assistant” can’t handle a basic returns scenario.

A Smarter Approach

Before bolting AI onto your ERP:

Start with the problem. Where are the inefficiencies, the delays, the guesswork?
Assess your data. AI is only as good as what you feed it—garbage in, garbage insights.
Test real-world scenarios. Not demos. Your workflows. Your users.
Measure ROI. If it doesn’t move the needle, it’s not worth it.

AI should amplify the ERP. Not to distract from the fundamentals of good process design and clean data.

Final Word

AI is powerful—but in ERP, it’s not a magic wand. It’s a tool. And like any tool, its value depends on where and how you use it.

So before adding that AI feature to your ERP wishlist, ask yourself: What’s the problem we’re solving? If you can’t answer that, you might just be looking at another solution in search of one.

Roger Pujol is a business improvement consultant and the founder of Champion Business Solutions, LLC. He speaks and writes about encounters, helping small to medium-sized businesses (SMBs) improve their business operations.