10 years of pentacor

The requirements for individual software solutions are constantly increasing – speed of development, adaptability and innovation are becoming key to differentiation in the market. With the rapid progress in AI, this trend is only accelerating. But how can an architecture remain flexible when technologies, teams, and processes keep changing?

In this talk, we show why continuous architectural work is the key to keeping pace with the dynamics of processes and technologies. We present our proven, practice-based approach, focusing on incremental modernization and domain-oriented system boundaries. Through continuous iterations, the approach enables systems to be modernized while still evolving in parallel.

Using real-world project examples, we make tangible how continuous architectural work creates sustainable flexibility – and how this secures the long-term innovative strength of our customers and their software solutions.

‍

Modern cloud organisations are faced with the question: How can we give development teams fast, autonomous access to cloud resources without compromising governance, security and compliance?

‍

The cloud operating model defines how responsibilities are distributed between platform teams and workload teams. The key challenge: Self-service and centralised control initially appear to be a contradiction. The presentation first introduces the cloud operating model and then shows how Crossplane not only enables this model, but also enforces it technically. As a Kubernetes-native control plane, Crossplane positions itself as the link between central governance and decentralised execution.

‍

The presentation demonstrates how Crossplane implements the operating model through abstraction, standardisation and control: platform teams define secure, reusable infrastructure patterns. Developer teams use these via self-service APIs – without being able to circumvent the specified standards. By combining declarative blueprints, automated validation and role-based access control, governance is transformed from an obstacle into an enabler. Using concrete examples, the presentation shows how organisational requirements are turned into technically enforceable patterns.

‍

RENGINE is a product name as a play on words from Recommendations and Engine. The system began as a license notification tool for digital vehicle extras. Over time it has grown into an engine that drives entirely new use cases and business value mainly in the area of digital vehicle extras.

‍

So what is the fuel that feeds this engine? We've all heard the phrase "Data is the new oil". We believe this premise is too narrow. For our product we see that knowledge-centric thinking was the real fuel to gain sustainable success.

‍

This presentation explores how this core philosophy has shaped the 9-year development journey of RENGINE from the ground up. Our approach was never focused solely on data collection. Instead, we constantly ask: What knowledge do we possess or generate that can not only improve our existing product but also serve as the foundation for entirely new use cases and scopes? For us, knowledge-centric thinking also means ensuring that the product reaches the right people ;-)

‍

Join us on this journey to see how this knowledge focus has driven a profound evolution within RENGINE – a development that transformed the entire system, our team setup, and, to be honest, ourselves as software engineers and visionaries.

‍

The menu of digital products has grown significantly over the past decade. An evolution in five courses – each building on the previous one, while expanding focus and reach.

‍

It starts with APIs for business functions, evolves into internal data products, grows further into external API offerings, expands through AI capabilities, ultimately leading to autonomous agents. With every stage, the audience widens: beginning with internal teams, extending to external partners, and ultimately including intelligent systems.

‍

But the right dish alone doesn’t make a successful restaurant. What happens in the kitchen? What ingredients are truly needed? And why do some fail in execution?

‍

The intriguing question: At which stage are you? And which combination matches your appetite?

Spoiler: The best strategy combines several courses into a well-designed menu.

‍

The Model Context Protocol (MCP) is transforming the way AI agents interact with enterprise systems and APIs. While traditional APIs were primarily designed for human developers, MCP creates a bridge that integrates AI models as full-fledged actors within enterprise landscapes. This enables them to directly access systems and become part of the system architecture.

‍

In this talk, we address the central question of how MCP differs from classical APIs, where parallels exist, and what implications this has for design, integration, and governance.

‍

Using a showcase, we demonstrate a concrete implementation: we show how existing APIs can be made accessible through MCP servers and present a holistic MCP blueprint. In doing so, we highlight how Azure cloud services can be leveraged to deploy, manage, and secure an MCP server.

‍

The talk provides practical insights from the development and use of MCP servers, including new challenges, solution approaches, and lessons learned.

‍

In modern retail enterprises, IT landscapes are increasingly decoupled along domain boundaries and organized according to business responsibilities – fully aligned with the principles of Domain-Driven Design (DDD) and “share nothing” architectures. For the evolution of products such as online shops, this brings clear advantages: new features can be delivered independently with reduced time-to-market. At the same time, real business value often emerges only by combining data across domain boundaries – for example, when newsletter sign-ups increase conversion rates or fraud detection improves through the integration of multiple data sources.

‍

But how can data-driven products be developed in such an architecture without violating the principles of autonomy and decoupling?

‍

This talk explores how the Data Mesh concept provides answers to this question and what impact it has on data architectures in decentralized enterprises. We will discuss how data can be treated as a product and made consumable across domains. Practical examples from different areas illustrate the full lifecycle of data-driven products – from identifying relevant data sources and building models to reintegrating insights into operational processes.

‍

From the perspective of a domain and product team, we share experiences and lessons learned in implementing such use cases. The goal is to provide actionable insights on how data-driven product development can be successfully realized in complex, federated IT landscapes.

‍

Companies possess valuable data and knowledge, which, however, are often difficult to access or not usable in the long term. With LinkedFactory, Fraunhofer IWU is presenting a data architecture concept that semantically structures production data, thus enabling digital twins. This creates transparency about processes, quality, and resources – and lays the foundation for assistance systems that simplify everyday factory operations.

‍

In the GENIUS research project, we are going a step further: Together with our partners at Fraunhofer IWU, we are testing methods to sustainably secure data and knowledge inventories, connect them with industrial production contexts, and simultaneously consider requirements such as the digital product passport (DPP). We are also relying on AI, particularly generative processes, that make knowledge specifically accessible and usable in the form of a digital factory buddy. The goal is to learn these concepts in production and prepare them in such a way that they can also be transferred to other industries. This creates a practical approach that helps companies preserve knowledge in the long term, make well-founded decisions, and strengthen their competitiveness.

‍

In our presentation, we would like to provide insights into how companies can use data and knowledge in a targeted manner in the future to better manage their own challenges.

‍

The role of software engineers is undergoing a fundamental shift: In a world where AI systems can not only generate code but also coordinate entire development processes, human responsibility, architectural thinking, and ethical reflection move to the forefront.

‍

In this talk, we explore what the future of software engineering looks like – not as a purely technical discipline, but as co-engineering between humans and AI. Using a practical example with the Responsible Vibe MCP Framework and the BMAD method, we demonstrate how AI can structure development flows, manage requirements, and provide context-sensitive suggestions – while humans act as architects, guardians of quality, and ultimate decision-makers.

‍

This raises inevitable questions: What skills will the software engineer of the future need? How do we ensure responsibility, transparency, and ethics remain intact? And how can we design tools that truly support teams instead of replacing them?

‍

Ten years of pentacor – ten years of innovation in software development. But the coming years will be different: AI is transforming not just how we write code, but how we collaborate, decide, and take responsibility. As AI tools generate code and orchestrate workflows, teams face new questions: What's the human role when machines can program? Where does our responsibility lie? What skills do we really need?

‍

In this panel discussion, experts from various industries share their perspectives on the future of software engineering. Together, we'll discuss how roles are changing, where AI helps, and where human judgment remains essential.

‍

Expect an open dialogue between practice and vision, opportunities and challenges – with space for difficult questions, bold predictions, and the opportunity to shape the future of software engineering together.

‍