Engineering Consultancy Services

🚢 How To Launch Big Complex Projects. How to reduce costs and schedule overruns, manage risks and be prepared for an unlucky turn of events ↓ 🤔 99.5% of big projects overrun budgets and schedules. 🤔 These are big relaunches, legacy re-dos, big initiatives. 🚫 Adding 15–20% buffer time/costs rarely saves them. ✅ Complex projects often follow “fat-tailed” distribution. ✅ There, overruns of 60–500% turn into big disasters. 🚫 Beware of unchecked optimism → unrealistic forecasts. 🚫 Beware of “cutting-edge” → untested technology spirals risk. 🚫 Beware of “bespoke/unique” → high chance of exploding costs. 🚫 Beware of “brand new team” → rely on tested and reliable teams. 🚫 Beware of “most advanced” → build small things, then compose. 🤔 The only way to prevent big disasters is to plan more. ✅ Best strategy: Think Slow (designing) + Act Fast (delivery). ✅ Good planning includes experiments, tests, simulations. ✅ Reference-class forecast → study mean actual cost, time. ✅ Track and review past projects in your company (cost, time). Things almost never go according to the plan — and on complex projects, they don’t even come close. We often assume that if we just thoroughly collect all the costs needed and estimate complexity or efforts, we should get a decent estimate of where we will eventually land. Nothing could be further from the truth. Complex projects have plenty of unknown unknowns. No matter how many risks and dependencies and upstream challenges we identify, there are many more we can’t even imagine. The best way to be more accurate is to define a realistic anchor — for time, costs and benefits — from similar projects done in the past. That’s what Prof. Bent Flyvbjerg calls reference-class forecasting (RCF) — experience-based, real-world outcomes that shape our estimates. No project is a snowflake; it always shares similarities with other projects. And however meticulous our calculations are, they usually approximate best-case-scenarios. Complex projects start with a deep deficit of experience. To increase the chances of success, we need to minimize the chance of mistakes even happening. That means trying to make the process as repetitive as possible — with smaller “work modules”, repeated by teams over and over again. It also means relying on reliable: from well-tested technology to stable teams that have worked well together in the past. And: always spend a bit more time planning, experimenting, testing and refining the plan before drawing a single pixel on the screen. It will pay off big time — every single time. I can only wholeheartedly recommend a book on “How Big Things Get Done” by Prof. Bent Flyvbjerg and Dan Gardner which goes in all the fine detail of how big project fails and when they succeed. It's not a book about design, but a fantastic book for designers who want to plan and estimate better. #ux #design

Back in 2022, we started to collaborate with Wallenius Marine AB to develop Oversea™ – a digital solution and fleet vessel support center-as-a-service. It collects and analyzes data from vessels around the world, enabling shore-based experts to deliver advanced decision-making support and tailored recommendations, such as optimizing routes based on real-time weather conditions. By combining ABB’s deep expertise in ship technology with Wallenius Marine’s operational know-how, we’ve created a powerful foundation for long-term impact across the maritime industry. Last year, we took our collaboration one step further and established a joint venture between ABB and Wallenius Marine. Together, we are working on delivering an even greater value to the maritime industry, helping customers not only perform better, but operate smarter and more sustainably for the long haul. Or, as we say in ABB, outrun – leaner and cleaner. Just recently, Jesper Lögdström, CEO of Oversea Fleet Support, joined our podcast to share real-world examples of how technology helps crews onboard and ashore make better decisions. Lots of great insights, have a listen! #EngineeredToOutrun https://lnkd.in/eiUVYF4n

🚨 “𝗪𝗲 𝗰𝗵𝗼𝘀𝗲 𝗕𝗿𝗼𝘄𝗻𝗳𝗶𝗲𝗹𝗱 𝗯𝗲𝗰𝗮𝘂𝘀𝗲 𝗶𝘁 𝘀𝗼𝘂𝗻𝗱𝗲𝗱 𝗰𝗵𝗲𝗮𝗽𝗲𝗿…” That’s what a client once told me. At first, the decision made sense. 💰 Lower cost. 🗓️ Shorter timelines. 🧩 Less disruption. But three months in… reality hit. They were buried in: ❌ Legacy custom code they didn’t understand ❌ Outdated processes holding back innovation ❌ Frustrated users who expected change but got the same old system in a new shell They had a new S/4HANA label - but nothing really changed. No smarter workflows. No process reimagination. No real value. And worst of all? They couldn’t explain to their leadership why they invested so much for so little transformation. That’s when it hit me: Clients don’t always “choose” Greenfield, Brownfield, or Bluefield. They default to it based on budget approvals and go-live deadlines. But they miss the dots they truly need to connect: 🔗 What’s broken in your current system? 🔗 What future do you want to build? 🔗 How much change can your people actually handle? Without those answers… Even the “𝗰𝗵𝗲𝗮𝗽𝗲𝘀𝘁” 𝗼𝗿 “𝗾𝘂𝗶𝗰𝗸𝗲𝘀𝘁” option becomes the costliest mistake. I stopped selling solutions. I started asking better questions: 👉 “Are you looking for innovation or just technical compliance?” 👉 “Do you want a new system or a new way of working?” 👉“What pain are you trying to remove and is your path really solving that?” And to make it easier, I built a one-pager that works in every workshop: 🏠 #Greenfield = Demolish and rebuild. Total reinvention. 🏠 #Brownfield = Renovate. Quick, but risky if you carry over junk. 🏠 #Bluefield = Selective remodel. Smart balance between change and control. SAP transformations aren’t about Green, Brown, or Blue. They’re about clarity. Clarity about what’s broken. Clarity about what success looks like. Clarity about how far you're willing to go to get there. Let’s stop being order-takers. Let’s be dot-connectors. Before recommending a path, dig deeper. Ask better questions. Paint the future not just the timeline. And if you need a conversation starter 📎 Use my Greenfield vs Brownfield vs Bluefield Cheat Sheet. It’s simple. Visual. Impactful. 🔁 Share it with clients. 🧠 Reflect before you architect. 🎯 Decide smarter and drive real transformation. 👇 Download. Share. Use. Because in SAP it’s never about the color. 😊 It’s always about the clarity. #SAPCareer #SAPSD #SAPSDTraining #SAPConsultant #CareerInSAP #SAPS4HANA #SAPJobs #SAPMentorship #SAPLife #SAPSuccess

I used Claude Code to build an agentic manufacturing AI troubleshooting assistant that reasons across a plant’s data to find a root cause (link to GitHub repo in the comments). The project contains >5,000 lines of code, all written by Claude. I have no coding experience. The project is built on top of a public time-series dataset of a real coal-fired industrial boiler in Zhejiang, China (link to the dataset and research paper in the comments). The assistant connects Claude to the following plant data sources through an MCP server (which contains 20 total tools): • a time-series database acting as a mock historian • a database of DCS alarm and event logs • a knowledge graph mapping relationships between equipment, instrumentation, control loops, and process areas • a RAG vector database containing plant documentation such as SOPs, troubleshooting guides, and equipment datasheets. Example use case: User asks: “Why is the boiler outlet steam temperature lower than normal today?” The agentic AI assistant works through a troubleshooting sequence on its own, leveraging the tools in the MCP server: 1. Queries the historian to confirm the deviation and define the abnormal time window. 2. Uses the knowledge graph to identify upstream equipment, instruments, and control loops that could influence outlet steam temperature. 3. Checks related process tags and alarm/event logs over the same window. 4. Narrows to a likely root cause, such as elevated spray water flow to the primary desuperheater. 5. Retrieves the relevant troubleshooting guide from the vector database and outputs recommended corrective actions. Important: An SME should always review the AI’s output for accuracy before taking action. I believe an AI troubleshooting assistant like this will exist in most chemical plants in 5-10 years. Some companies are probably ahead of the curve and are already designing something similar. If implemented correctly, annual troubleshooting hours and downtime could decrease significantly, and engineers will spend more time improving their processes rather than firefighting. Keep in mind that this is a small prototype and implementing in a real chemical plant will be far more complex. IT barriers alone would be a nightmare. These problems are solvable though. I believe the end goal is worth it. Please reach out if you are working on something similar or would like to chat more about the project. What improvements would you make to this design?

AI didn’t assist engineers here. It designed the rocket engine. What do you think? LEAP 71 just proved something big for engineering and AI: • A liquid rocket engine was autonomously designed by a physics-based AI system (Noyron) • 3D-printed as a single copper part • Hot-fired successfully on the very first test • No traditional CAD, no manual iteration loops This wasn’t trial-and-error. It was pure physics + computation + manufacturing constraints encoded in software. Once the model exists, new engine variants can be generated in minutes, not months. Why this matters: Rocket engines are among the hardest machines humans build: • ~3,000°C combustion temperatures • Cryogenic propellants • Extreme pressure, vibration, and thermal stress And yet… the first design worked. This isn’t “AI will replace engineers.” This is engineering moving from drawing to defining intent — and letting computation do the rest. Same shift we’re seeing in: • Semiconductors • AI infrastructure • Advanced manufacturing • Robotics & simulation Design is becoming software. Testing is becoming data. Iteration speed is becoming the real advantage. The future of engineering just fired on a test stand 🚀 #AI via @codeintellectus and Joel Gomes #Engineering #Aerospace #ComputationalDesign #AdvancedManufacturing #3DPrinting