The rapid development of generative artificial intelligence (GenAI) and large language models (LLMs) is on the verge of fundamentally changing software architecture. In this talk, we will take a “look into the crystal ball” to explore what impact these technologies will have on traditional architectural decisions, code lifecycle and human-machine interaction.

We will explore how the use of LLMs could lead to traditional architectural issues such as choice of programming language becoming less important, while new challenges arise in the areas of security and ethical responsibility. We will also discuss the potential benefits of more automated and optimized development practices enabled by these technologies.

This talk is aimed at software architects, developers and technology enthusiasts who want to understand how GenAI and LLMs will shape the future architecture of our software landscapes. Together we will take an informed look at the future of software development and discuss how we as professionals can prepare for it.

This recording from the Software Architecture Gathering 2025 was powered by iSAQB e.V. in collaboration with Software-Architektur im Stream.

Many thanks to Kameramensch-Filmproduktion https://kameramensch.de/: Johannes Klemt, Nico Schallat, Vladimir Traut for providing the footage.

Event tech by Team Kingsize https://www.team-kingsize.de/

Follow iSAQB:

• Instagram https://www.instagram.com/isaqb.official/

• LinkedIn https://www.linkedin.com/company/isaqb.org/

• Mastodon https://mastodon.social/@isaqb

• BlueSky https://isaqb.bsky.social

• Facebook https://facebook.isaqb.org

At the recent Software Architecture Gathering, Eberhard gave a talk on models, modularization, and Bounded Contexts, while Alistair explored similar themes in conversations with attendees. In this episode, the two come together to discuss the key concepts behind these topics, compare their perspectives, and answer questions from the audience.

Here is the quote from Parnas’ paper that Alistair referred to:

“We propose instead that one begins with a list of difficult design decisions or design decisions which are likely to change. Each module is then designed to hide such a decision from the others.”

• David Parnas: On the Criteria To Be Used in Decomposing Systems into Modules

• Alistair Cockburn: Design in Object Technology 2: The Annotated Class of 1994 EPUB Printed

• Residuality Theory with Barry O’Reilly

Stimmen aus der Praxis – Live-Stream mit Erfahrungsberichten und Diskussion

Wie erleben Menschen, die in der IT unterrepräsentiert sind, ihren Arbeitsalltag wirklich? In diesem Live-Stream bringen wir ungefilterte Stimmen direkt auf die Bühne der IT-Tage.

Im Vorfeld haben wir persönliche Erfahrungsberichte gesammelt – von Mikroaggressionen im Team-Meeting bis zu strukturellen Hürden beim Karrierestart. Diese anonymisierten Geschichten lesen wir vor und diskutieren gemeinsam: Was läuft schief? Was funktioniert bereits gut? Und vor allem: Was kannst Du selbst, Dein Team oder Deine Organisation morgen konkret anders machen?

Aus der Praxis für die Praxis: Diese Session richtet sich an alle, die Diversität nicht nur als HR-Thema verstehen, sondern als strategischen Erfolgsfaktor für ihre IT-Organisation. Ob CTO, Teamlead oder Entwickler:in – hier erhältst Du ungefilterte Einblicke aus dem Projektalltag, die sonst oft unsichtbar bleiben, und kannst konkrete Impulse direkt in Deinen Alltag mitnehmen.

Was Du mitnimmst: Echte Erfahrungen statt Theorie, Verständnis für Barrieren im IT-Alltag und Handlungsfelder für inklusive Teams – so konkret wie möglich, so anonym wie nötig.

Mit Liam Bergh, Sabine Wojcieszak und Eberhard Wolff

In this engaging live fishbowl session from the Software Architecture Gathering, Vaughn Vernon, Cheryl Hung, Avraham Poupko, Eberhard Wolff, and Ralf D. Müller tackle one of the most pressing questions in the field: Is software architecture about human intelligence or artificial intelligence? As AI tools increasingly design systems, analyze code, and critique architectural decisions, the panel debates whether these technologies augment or replace the architect’s role. They explore the nuanced balance between machine-generated patterns and human creativity, the ethical and accountability challenges of AI-driven architecture, and practical ways architects can thrive in an AI-augmented future. Audience participation ensures a lively, thought-provoking dialogue on the evolving craft of software architecture.

Livestream production by ‪Kameramensch-Filmproduktion‬: Johannes Klemt, Nico Schallat, Vladimir Traut

Event tech by Team Kingsize

Follow iSAQB:

• Instagram

• LinkedIn

• Mastodon

• BlueSky

• Facebook

In this special live episode from the Software Architecture Gathering, Clare Sudbery joins Eberhard for an open conversation about learning, empathy, and psychological safety in tech. Building on her talk “Let’s Stop Making Each Other Feel Stupid”, Clare reflects on how the culture of expertise in software development can unintentionally discourage curiosity and collaboration. Together, they explore practical ways to create environments where questions are welcomed, mistakes become opportunities for growth, and knowledge sharing thrives behind the scenes of architecture and development.

Links from Clare’s talk

In this live episode from the Software Architecture Gathering, Claudine Allen joins Ralf to explore how software architects can leverage large language models (LLMs) as collaborative design partners. Drawing from her research, case studies, and experiments, Claudine discusses a practical framework for engaging with LLMs throughout the architecture analysis and design process. They delve into key questions such as what to tell LLMs, how to prompt them effectively, and how to assess risks—highlighting the enduring importance of human judgment in architecture while embracing the transformative potential of AI-powered collaboration.

In this special live episode from the Software Architecture Gathering, Jacqui Read joins Eberhard and Ralf for an insightful conversation on the rise of diagrams-as-code and the role of AI in software documentation. Jacqui highlights how using textual notations like PlantUML, Mermaid, and Structurizr transforms diagramming from an artistic task into a systematic, data-driven process. They discuss the benefits, limitations, and need for critical understanding when working with AI-generated diagrams, empowering teams to improve consistency, collaboration, and automation behind the scenes of architecture workflows.

Digitale Souveränität ist aktuell ein sehr heiß diskutiertes Thema und im Mainstream der Gesellschaft angekommen. In dieser Episode diskutiert Sven Müller mit Eberhard Wolff den Einfluss von digitaler Souveränität auf Software-Architektur. Ziel ist es, die digitale Handlungsfähigkeit unserer Organisationen zu stärken. Aspekte sind nicht nur die häufig diskutierte Unabhängigkeit von US-Cloud-Anbietern, sondern vor allem gezielte Make-or-buy-Enscheidungen, der Einsatz von offenen Standards oder Open Source nicht nur zum Sparen von Lizenzkosten.

Links

• Whitepaper Digitale Souveränität im Software Engineering

• Software Architecture Gathering 15% Rabatt mit Code SATV_SAG2515