Post by Rick Bouter

Siemens Realize ‘26 day 2: Clean and stabilise the core, embed AI and integrate around it The problem in engineering organizations: legacy PLM, ERP, and shop-floor systems accumulate heavy customization over decades, fragmenting design, BOM, and production data across siloed tools. The goal for engineering organisations is to increase engineer productivity, quality, and velocity on stable platforms that can scale towards future demands. 1️⃣ Diagnose before building | Treat root causes, not symptoms, every fix and customization decision starts from understanding the actual problem, not patching around it. 2️⃣ Stabilise the foundation | Do not build more customizations on weak or unstable foundations. Fix the architecture first. 3️⃣ Govern customization, do not ban it | Customizations are not wrong, but each one should be a deliberate trade off weighed against available standard solutions and strategic direction. 4️⃣ Design for loose coupling | Move away from tightly coupled systems toward loosely coupled modules, in a layered architecture with clear objectives per layer. 5️⃣ Make change impact visible, keep humans accountable | A developer changing part of the code is not a standalone action, it can impact deep into modules, libraries, and classes, so insight into change impact is needed. AI can support this analysis, but AI cannot take over human sign off, the human stays in the lead and remains responsible. Thanks Raghu Timmanagoudar and Sagar Chaudhari , TOGAF® on their inspiring pitch: Code quality and decustomization to future proof PLM. A great and ongoing journey. With Accenture team Martijn, Mark, Daan, Eduard, Nitin, Srikiran, Kapil, Csaba, Adela, Niranjan, Nicole, Lucija, Victor, Antti, Tim, Götz, Mark, Cleodene #Siemens #SRL #SiemensRealizeLive #RealizeLive #Accenture #engineering #engineeringai #digitalthread #manufacturing #cleancore #decustomization Accenture the Netherlands