Enterprise AI & Decision Architecture
Developing practical approaches that connect evidence, semantics, access and governance in SAP-heavy environments.
RomanUnterstögerAbout Roman
From a business question
to a working system.
I clarify what a company means by a question, how the answer is calculated and what an AI system is allowed to do with it.
The technology changed. The task behind it did not.
Whether the result is software, reporting or AI, business intent must become precise enough for a system to act on it reliably. That translation is my work.
I write about problems that show up in real projects: conflicting metrics, unclear ownership, missing evidence and AI systems that answer anyway.
In practice
Workshops quickly reveal where business, data and technology mean different things. That is where my work begins.
Business intent into decision logic
Turning management questions into KPIs, definitions and traceable reporting systems — across the boundary between business and data.
Digital products and processes
Designing custom software, digitizing operational processes and learning what adoption requires beyond a technically correct system.
Software and change
Engineering systems, coordinating technology and managing change and releases in complex enterprise environments.
Technical foundation
Certifications.
BTP, Data Fabric, BW/4HANA and Generative AI.
Working principles
Start with the evidence.
An answer does not become correct by sounding convincing.
Put meaning in the system.
Definitions and rules need to be visible and versioned.
Ask when the question is unclear.
A good follow-up question often prevents a bad decision.
Name the person in charge.
Responsibility only works when it comes with authority.
More on LinkedIn.
That is where I publish new articles and shorter notes from my work with Enterprise AI, SAP and analytics.
Follow on LinkedIn ↗