A couple of firsts

It’s a sunny but cold fall day in Mannheim, and with the November release still around the corner, this episode is all about finding answers. In this in-between episode of the AntiPattern Podcast, Christian, CTO, and Michael, CEO of Elevating Patterns, take on the questions you sent them via social media, email, and at events. First up: are they afraid of AI? Not at all. Without AI, they couldn’t have built the company the way they did with such a small team. Many of the scary headlines, they argue, are part of the hype cycle and a good portion of marketing. Their own experience looks different: guardrails are working, and the latest models even ask for explicit confirmation before running risky shell commands. What actually worries them more are hallucinations and non-deterministic answers. Which leads straight to the next question: what’s your moat? The answer is the EP Core. Built over two years, it handles configuration and lifecycle management, and it can’t be changed by AI. Christian and Michael love vibe coding and want customers to vibe-code on top of their platform, just not the parts where you need deterministic results and a reliable audit trail. Because when it comes to critical product and process data, “if you’re lucky, it works” isn’t good enough. They also describe their dream customer: a problem-aware company, ideally from the German machinery Mittelstand, that’s fed up with its current PLM vendor, understands its own processes, and believes the process should lead while the tool follows. Bonus points for being open to working with a startup and ready to break up silos between PLM, ERP, and MES. And will Elevating Patterns build CAD connectors? Short answer: no. CAD data management is a solved problem. Keep your PDM, connect it to the digital thread, and leave the desktop connectors to partners. The highlight of the week: a sneak peek at the next release, with drag and drop across browser tabs and windows. Something many big vendors still don’t support, even though the browser API has been around for years. Built with AI in about 20 minutes. Plus: Elevating Patterns now has a community area where you can register for events, vote on features on the roadmap, and leave your feedback. Got a question of your own? Send it in, and you might find the answer in a future episode.

It’s a cloudy day in Mannheim, and the perfect segue into today’s topic. In this episode of the AntiPattern Podcast, Christian, CTO, and Michael, CEO of Elevating Patterns, talk about what cloud really means, and why there is still a lot of misunderstanding about it in the market. They start with a key distinction: cloudified vs. cloud native. Putting yesterday’s desktop application into a single Docker container and hosting it per tenant doesn’t make it cloud. Cloud native means building for the cloud age from the start: microservices, independent scaling, and containers that know how to start, register, and say hello to their neighbors. From there, they take you along on their infrastructure journey. From Docker Compose to lightweight orchestration with K3s, and eventually to full Kubernetes. Which raises the big question: run Kubernetes yourself, or go for managed Kubernetes with one of the hyperscalers? The heart of the episode is the upcoming Cloud Console. It lets customers book their own demo instances, manage tenants, environments, and dev, test, and production instances, and handle subscriptions, all in self-service. The ambitious goal: a platform so easy to use that you never have to talk to anyone. Christian and Michael also share how customizations, apps, AI agents, and data move between instances via packages and namespaces, and what a master’s thesis taught them along the way. And of course, there’s AI. With a recent AI-built remake of an entire office suite making headlines, they discuss what software is actually worth rebuilding, and why it’s usually not the vendor’s product. Introducing a new ending, the anti-pattern of the day: software vendors don’t need to fear being copied. What’s really worth rebuilding is the custom code customers had to add because the standard software couldn’t do the job. Plus: Elevating Patterns has a new website, with a community area and a public roadmap you can help shape.

What makes a release more than just a list of new features? In this episode of the AntiPattern Podcast, Christian, CTO, and Michael, CEO of Elevating Patterns, take a closer look at EP Digital Thread Platform Release 2608 — the August 2026 release packed with capabilities designed to make working with complex industrial data more powerful, flexible, and collaborative. They walk through the highlights of the release, including full-text search directly in the command bar, a new GraphQL endpoint, and configurable versioning schemes that allow teams to tag commits of a dataset with meaningful version numbers. One feature in particular could be a game changer for mechanical engineering: Branch & Merge. Inspired by the way developers work with Git, it brings branching concepts into the world of engineering data and opens up completely new ways of collaborating on variants and changes. The AI experience is also taking a major step forward with an enhanced AI Chatbot featuring Ask, Planning, and Agent modes. And with server-side events, users can now see what others are doing in the system in real time — making collaboration much more transparent. Last but not least, Edit from Grid makes working with data faster and more intuitive directly where users see it. Beyond the release itself, Christian and Michael also share an update on the latest Variability Research PoC and what the team is learning along the way. Release 2608 is more than another release. It’s a glimpse at where the Digital Thread is heading.

For one day, startups, universities, and research institutes came together in Mannheim to talk about Industrial AI. Sounds like the beginning of a joke—or perhaps the beginning of a fellowship. In this episode, Christian and Michael look back at the AI Research Day, which they co-hosted with SysDICE. The big question: What does it actually take to make AI work in complex industrial environments? One thing quickly became clear: Industrial AI is all about context. Models need more than data—they need to understand the processes, systems, and technical relationships behind it. Model harnesses might become a crucial piece of the puzzle, while reference architectures could provide the structure needed to make all of this work together. Yes, we know: reference architectures can get a little academic. But maybe this time, it’s worth the risk. The most concrete outcome of the day? The participants decided to write a joint whitepaper on Industrial AI. With one central idea at its heart: the Digital Thread as the key to giving AI the context it needs to understand deep technical and industrial processes. So, is this the beginning of a new fellowship? Or just another anti-pattern waiting to happen? Let’s find out.

In this episode, Christian and Michael take a deep dive into the latest release of their Digital Thread Platform: version 2605. Packed with new capabilities and long-awaited enhancements, this release marks another major step forward for the platform. They walk through some of the highlights, including Single Sign-On support, upgradable EntityTypes, configurable dashboards, and extended virtual properties that make data modeling even more flexible and efficient. The star of the release, however, is the new AI Data Modeler. Christian and Michael discuss how AI is becoming an integral part of the platform experience, helping users accelerate modeling tasks and unlock value from their digital thread faster than ever before. Join them for a tour through the new features, the ideas behind them, and a few stories from the journey of bringing release 2605 to life. Because sometimes, if you can’t release it, maybe the release should release you.

In this episode, everything revolves around people, collaboration, and partnerships. Because in the end, it is all about people — the people we meet, collaborate with, and build products together with. We share news about our successful pre-seed financing round and welcome our new business angel Oliver Loose, who not only invests in the company but also brings extensive SAP experience and valuable industry insights. We also talk about two academic projects we are currently involved in: Etienne’s master thesis focuses on concepts for one-click deployment and the architectural implications for our platform. At the same time, Nico Kasper has just submitted his PhD thesis, which Michael has helped supervise over the last couple of years. Most importantly, we discuss our first official technology partnership with RheoNet and Sebastian Behrend. Together, we are building the first partner application for material flow simulation on our platform. Along the way, we are learning what capabilities a platform really needs in order to enable meaningful partner ecosystems and third-party applications.

In this episode of the Anti-Pattern Podcast, Christian and Michael take you Back to AI: Lessons from the Future — reflecting on three key insights from the prostep ivip symposium. First, they explore why “SaaS is dead” is more than just a provocative statement. It’s not about how software is delivered anymore, but about how easily it can be customized. As AI enables the creation of highly tailored solutions at low cost, static, out-of-the-box software is losing relevance fast. Second, they discuss AI as a utility and the big open question: who will capture the value? Will it be the LLM providers, or those who successfully embed AI as a core architectural building block into their platforms? Christian and Michael share their perspective — and why platform players might have a strong position in this evolving landscape. Finally, they dive into the human factor. What does it actually mean to work with AI? How do we deal with the cognitive load, the disruption, and the changing role of humans in increasingly AI-driven systems? Three lessons, one conversation — and a look at what might shape the future of software and work.

What happens when enterprise software stops being configured in code—and starts being shaped by prompts? In this episode, we explore how AI is fundamentally changing the economics of customization. If tools can be configured through prompting, the cost of tailoring enterprise IT could drop dramatically—Michael estimates by as much as 90%. We also share what we’re building at Elevating Patterns: • Connectors to external systems that can generate themselves • Data models that are created directly from Claude The result? A shift from manual integration work to AI-driven composition—faster, cheaper, and far more flexible. Is this the end of traditional customization as we know it? #AI #EnterpriseSoftware #NoCode #Integration #DigitalTransformation

In this episode, we're sharing a few important milestones from behind the scenes at Elevating Patterns — first successes, first learning moments and what drives us forward. For the first time, we are in the process of closing an external financing round — and we're excited to welcome our first business angel on board. We talk about what this step means for us, why we decided to take this route, and how it will support the next phase of our growth. Another first: we've made a new hire. And yes, that officially triggered the launch of our brand-new "new hire alarm." We share who's joining, what this means for the team, and how we're thinking about building Elevating Patterns step by step. And what that has to do with Steve Ballmer the former CEO of Microsoft. Finally, we let you know where you can meet us live over the next weeks: • Prostep iViP Symposium, April 14–15 in Frankfurt, Germany • Threaded! at Develop3Dlive, March 25 in Warwick, UK If you'll be there — come say hello. We're always happy to connect with fellow pattern thinkers, system shapers, and anti-pattern spotters.

In this episode of the AntiPattern Podcast, we explore the shift from a passive to an active Digital Thread — and why this distinction makes all the difference in modern product development. We discuss the two fundamental integration challenges that continue to block real progress: horizontal integration in a joint data model and vertical integration connecting multiple tools — be it in the cloud or on-prem. While these challenges are often treated as inevitable limitations, we make one thing clear: integration across tool boundaries will not fail because of Elevating Patterns. We also address a common misconception head-on. Elevating Patterns is not building yet another data analytics tool. An active Digital Thread requires more than semantics and dashboards — it demands data orchestration, robust data management, and engineering-grade concepts such as versioning, variability, and change management. Moving from passive data to active systems means enabling products to react, adapt, and evolve across their lifecycle. That transition — from observation to action — is what truly defines an active Digital Thread, and it is exactly where Elevating Patterns sets itself apart.