Let's be brutally honest. In recent meetings of many boards around the world, panic and greed have been disguised as «strategy». The Artificial Intelligence in SAP has become the new corporate fetish. Someone, probably with more enthusiasm for reading trend headlines than knowledge about the architecture of their own systems, pounded the table demanding a «Copilot», a «Joule» or a «corporate ChatGPT».
They want ERP to magically talk to them, predict the future and tell them how to save next year's EBITDA. It's the shiny new toy. It's the promise of ultimate automation.
But I have news that your consultants don't want to give you, because selling subscription licenses is infinitely more lucrative than telling the uncomfortable truth: If you try to deploy Artificial Intelligence in SAP about your current data mess, you're not going to get business insights. You're going to get a 4K X-ray of your own operational mediocrity.
AI doesn't come to save you; it comes to expose you.
Artificial Intelligence in SAP and the mud of your data
The problem is not the technology. Language models (LLMs) are impressive tools, as confirmed by technology trend experts at Forbes: Why AI Pilots Fail. The real problem is the rotten substrate on which they intend to operate.
Imagine this scenario, which you are probably familiar with: A CEO asks his brand new assistant connected to the ERP: > «Analyze the margin drop in the Manufacturing line this quarter and compare it to the immobilized stock.»
The executive expects a Wall Street-type response, a fancy paragraph justifying complex decisions. But the machine has no political tact or fear of being fired. If the Artificial Intelligence in SAP were honest, the answer would be:
«I can't calculate the actual margin. The 40% of import costs are charged to a generic accounting account called ‘Other Expenses’ because no one in Purchasing wanted to create the correct service items. Also, your theoretical stock does not match reality because the last physical inventory was eight months ago. Your data says we're profitable, but your cash flow says we're in crisis.»
This is what I call Corporate Hallucination. We often say that AIs «hallucinate» when they invent data. But in the business world, the delusion is human. It is the company that has been lying to itself for years, and technology has only amplified that delusion.
The «Silent Cancer» of Excel and Cowardice
This is where we go deeper into the wound. The reason why the Artificial Intelligence in SAP is going to fail miserably in the 90% of organizations is not a software failure, it is a matter of culture and cowardice.
Over the years, we have allowed and normalized what I call the «Parallel Excel Cancer«, a subject I have discussed in depth in my previous articles.
The user does not load the correct information into the system because «it is slow» or «rigid». So, he creates his own reality in a spreadsheet. The boss validates that sheet. The board is presented with results based on that sheet. The central system is left empty of truth, turned into a simple fiscal repository.
When you connect an AI, it ignores the personal Excels and goes to the single source. And what does it find?
- Duplicate Customers: C1001, C1002, C_Final Who are we really selling to?
- Garbage items: Descriptions such as «Z-RODS» or «TEST».
- False Dates: Deliveries recorded on the last day of the month only to meet the administrative KPI.
If your data is garbage, your AI will be an automatic dispenser of strategic errors at the speed of light.
The end of «Intuition» as an excuse
There is an underlying fear that no one admits out loud, but which is the real barrier to adopting the Artificial Intelligence in SAP.
Many managers have built their careers on «intuition» and «business acumen». For decades, that was a fancy way of saying: «I don't have reliable data, so I'll guess and use my charisma to sell the idea.».
AI threatens that status. Technology democratizes access to information (if it exists) and eliminates the mysticism of the leader who «feels» the market.
A well-implemented system not only answers questions, it challenges assumptions. I could tell the business manager:
«Your discount campaign didn't increase loyalty, it just destroyed margin on products that were already selling themselves.».
That's frightening for an executive who has always navigated ambiguity. The absolute transparency that AI demands is the natural enemy of traditional office politics.
Do you want AI? Clean the blood first
Don't spend a dime on licenses if you're not willing to do the dirty work first. For implementation to work, you need to get back to the basics of ethics and governance:
- Forensic Audit of Master Data: Stop creating patches. Clean up your item master and business partners as if the viability of your company depended on it.
- Kills Shadow IT: If the data is not in the ERP, it does not exist. Period.
- Real Accountability: The typing error is no longer «an administrative detail», as I mentioned in my reflections on Digital accountability . In this new era, a wrong input is a wrong strategic decision automatically made by a machine.
Conclusion: The mirror is not to blame
Technology is the ultimate mirror. Many IT managers and consultants are terrified, not because algorithms are taking away their jobs, but because the Artificial Intelligence in SAP will demonstrate that, during the last ten years, they were not managing information, but administering chaos.
Next time you ask for «innovation» and «digital transformation,» ask yourself: Are you ready for a machine to tell you the brutal truth about how your company really operates, or would you rather continue living in the sweet lie of your parallel data?
The choice is yours. Truth hurts, but obsolescence kills.
