Here’s a quick rundown of what’s happening: SAP has acquired an 18-month-old German AI lab for $1.16 billion. Not a chatbot startup, not a GPT wrapper. Prior Labs builds Tabular Foundation Models, transformer models that understand rows, columns, and tables. Exactly the kind of data our customers rely on: purchase orders, accounts receivable, inventory levels, energy flows, and customer contracts.
In other words: SAP isn’t buying a language model. SAP is buying the ability to understand its real gold, the business data that has been sitting in its systems for decades.
Why this matters to us at SUPERP
SUPERP was founded in 1999 out of frustration with SAP implementations that were falling apart at the seams. Our founders took a holistic view of the entire process, from implementation to management. That has always been our approach: stepping in where things get complicated, where SAP, processes, and business logic converge.
And that is exactly what this acquisition addresses. Because what SAP is trying to do now, deriving meaning from tables full of business data is not a technological issue. It is a process issue. We see this every day with network operators migrating to S/4HANA, retailers looking to take quality engineering to the next level, and manufacturers seeking to connect their disparate systems.
An AI that “understands” rows and columns sounds wonderful. But if the master data is messy, if processes are flawed, or if the S/4HANA environment is out of whack, that AI will only spew out sophisticated nonsense. Garbage in, smart-looking garbage out.
From a system of record to a system of intelligence
| Today: System of record | Tomorrow: System of intelligence |
| SAP saves the transaction. A person reviews it. A person makes a decision. A person carries out the next step. | SAP understands the transaction. It predicts the next one. An agent suggests the action. A specialist makes the decision with more context. |
Note the last line: a professional makes the final call. Because that, to me, is the crux of the matter: AI does not replace expertise. It shifts where expertise is needed. No longer in manual execution, but in calibrating, validating, and critically evaluating what the models propose. Someone has to understand why that prediction is correct or why it is dangerously wrong.
What we take away from this and what we do
At SUPlabs, our co-innovation platform, we’ve been working with clients on these kinds of challenges for years. Not as a passing fad, but as concrete steps: business cases, pilots, and proofs of concept. We don’t do it for our clients, but with them. That’s not just a marketing message; it’s how you prevent an AI investment from becoming nothing more than an expensive PowerPoint presentation.
We also eat our own dog food. Since 2022, our back office has been running on mijn.SUPERP, a platform we built entirely in-house using SAP UI5 and Node.js, because the standard tools didn’t align with how we work. The same principle applies to AI: you only truly learn it by applying it to your own processes.
So here are three things I’m sharing with clients now that this deal is in place:
- Invest in your foundation first. An AI layer built on top of a shaky S/4HANA environment won’t deliver insights, it’ll just create problems on a massive scale. Master data, integration, and process discipline come before the AI party.
- Think in terms of outcomes, not licenses. The question isn’t “Which SAP AI module should I buy?”, but “Which decision do I want to make better, faster, or more automatically?” You work backward from that to the technology, not the other way around.
- Keep expertise in-house. The smarter the models, the more you need people who can interpret, challenge, and improve the results. That’s not something you can simply purchase as a standalone service; it’s something you build together.
In conclusion
SAP is investing $1.16 billion to make the transition from “storing” to “understanding.” That’s no small shift. But the value of that shift isn’t in the press release it’s in what our customers will do with it. And that’s what we’re here for.
