Recap | CMS Summit 2026 in Frankfurt

What could be more fitting for the CMS Summit than a venue where everything revolves around communication? On 12 and 13 May, CMS experts from a wide range of fields gathered at the Museum of Communication in Frankfurt to discuss digital platforms, AI, content management and collaboration. The venue itself was a perfect match for the character of the event: open, versatile and brimming with lively discussion!

Two days packed with expert insight

The CMS Summit 2026 was far more than a traditional conference featuring a series of talks. Between sessions, round tables, group discussions and spontaneous conversations in the corridors, the event fostered exactly the kind of exchange that is often lacking in day-to-day working life. 
Instead of buzzwords and promises of the future, the focus was on challenges: How do teams work together efficiently? How is AI changing the way we communicate? And how do you create digital experiences that are truly inclusive? 

Anika, Emina, Søren and Olivier from our dkd team were there, accompanied by one of our clients. We were also present as dkd, acting as a sponsor of the event. And even though the programme was packed, we actually managed to find a little time in between to wander through the museum’s exhibitions. Amidst old telephone systems and historical means of communication, it quickly became clear that the way people communicate is constantly changing.
Some challenges, however, remain surprisingly similar – only today they take place via messaging apps rather than fax machines. You can find an overview of the talks here! 

EI rather than AI: Why collaboration is the real scaling factor

Technologies are becoming faster, processes more complex, and teams are increasingly working across different disciplines. This makes the question of why projects actually fail all the more interesting. Florian Keitgen, Head of Product Development at b13, addressed precisely this question in his session ‘The Hidden Scaling Risk: When the Human Layer Becomes the Bottleneck’. 

Rather than looking at systems or technical challenges, the presentation deliberately focused on the human factor. Using open-source communities as an example, it was clearly demonstrated how quickly projects can fail due to dynamics that, at first glance, have little to do with technology. Unclear responsibilities, communication via individual people, or invisible contributions often have a significantly greater impact on collaboration and speed than technical hurdles. 

Particularly fascinating was the focus on issues that arise in almost every working environment but are rarely addressed openly. Status-seeking behaviour, gatekeeping and informal power structures often creep in gradually and, particularly in more complex organisations, quickly become real obstacles. Perhaps the most interesting insight from this session, however, was that amidst all the discussions surrounding AI, a completely different topic suddenly came to the fore: EI, or emotional intelligence.
For the more complex digital projects become, the more important it is to be able to clearly define responsibilities, consciously shape communication and build stable, long-term collaboration. Perhaps sometimes we don’t need more AI, but rather a little more EI first. 

When the Internet Talks Back  

The debate surrounding large language models often swings between two extremes. Either AI is portrayed as the solution to almost everything, or as a major threat to existing ways of working. Liz Nelson, Vice President of Product and Technology at Sitecore, offered a much more nuanced perspective in her session ‘The Internet Talks Back‘. 

The focus was on a statement that sounds surprisingly simple at first: large language models are essentially the internet in condensed form, answering questions. It was precisely this perspective that suddenly made many connections more tangible. Particularly fascinating was the glimpse behind the actual response of an AI system. For before a result becomes visible, numerous processes and so-called hidden layers run in the background, determining how content is processed and answers are compiled. 

One example in particular stuck in the mind: the comparison between human writing and AI-assisted writing. Human texts often thrive on a variety of voices, nuances and individual styles of expression. When, on the other hand, content is created on a large scale with the help of AI, linguistic uniformity quickly sets in. It was a thought that probably prompted some in the room to briefly reflect on their most recent LinkedIn posts. 

Alongside the technical insights, the focus was primarily on a practical question: how can AI be used effectively without losing one’s own brand voice and individuality? 

The approaches presented were deliberately pragmatic. Start with a clear strategy, manage processes, keep people involved in the decision-making process, and continuously refine results. Because even though AI can speed up processes, good communication still doesn’t happen automatically at the touch of a button. 

Interested in a GDPR-compliant AI solution? Starfruit AI is the answer! 

Digital inequalities often arise sooner than we think

At first glance, digital products often appear neutral. Content is displayed, processes run, and users navigate through an application. At the same time, behind every platform lies a multitude of small decisions that influence how people experience digital offerings. This was precisely the topic addressed by Antonia Fedder, Designer for impact-driven brands, in her session "Where digital bias hides: configuration, content, and communication". 

The focus was not only on the content itself, but also on configurations, content models, workflows and communication patterns. This is because every decision in digital projects automatically carries with it assumptions about who an offering is being developed for and who may not have been sufficiently taken into account. 

This is precisely why the session was particularly engaging: digital bias often doesn’t just arise in major strategic decisions, but begins in small, everyday processes. It is often seemingly minor details that later have an impact on the user experience. What was particularly successful was that the session not only highlighted challenges, but also generated specific questions that can be directly applied to projects or workshops.

The subsequent round-table discussions also explored these topics in greater depth. Among other things, they addressed the question of how reliable automated GDPR tests actually are, and what impact new developments such as the Digital Fairness Act might have on digital products in the future. Issues such as artificial scarcity and mechanisms that reinforce addiction clearly demonstrate that digital responsibility must be considered in an increasingly broader context.

One key insight ran through the entire session: inclusive digital experiences do not happen by chance. They come about when teams consciously question the assumptions that are already built into platforms, content and processes.

Do you have any questions about accessible websites or the latest requirements set out in the BFSG? Our experts from the dkd team will be happy to help.

Think content first – not just before launch

In many projects, the process follows a similar pattern: discovery, wireframes, design, development, and then, at some point towards the end, the content is added. It was precisely this sequence that was deliberately scrutinised in the session "WYMIWYG: What You Model Is What You Get" by Markus Schork, Director of Technology Solutions at Codal.

The starting point was a question that is actually surprisingly commonplace: why do even small changes to a website often take longer than they should? Using concrete examples, it became clear that the cause often lies not in the CMS itself, but in the way content is modelled.

This is because systems that promise maximum flexibility often have the opposite effect in practice. Complex structures, unnecessary dependencies and extra steps in day-to-day editorial work quickly mean that, whilst content may be flexible in theory, it becomes unnecessarily time-consuming in practice. A particularly interesting idea was to consider content much earlier in projects – not just after design and development, but immediately following the discovery phase. After all, content forms the basis of what users ultimately see, read and experience.

One insight ran through the entire presentation: a website project does not begin with the go-live. People often underestimate how many decisions, structures and processes long before that point influence its later success. Start with content, not design.

AI-ready content begins long before the first prompt

Content is currently undergoing a noticeable transformation, not only at the level of individual texts, but above all at a structural level. It was precisely this shift that Kate Kenyon, Head of Content Design, Operations & Localisation at JPMorgan Chase, addressed in her session “The hidden work behind AI-ready content”. 

The focus was less on which new tools are currently available and more on what conditions need to be created to ensure content works in the long term. One statement in particular stuck in the mind: "Metadata is AI's love language." There is a very practical reality behind this statement. The better content is structured, described and provided with context, the easier it is for AI systems to understand, categorise and reuse it.

It was also fascinating to see how roles within teams are changing. Whilst certain traditional tasks can increasingly be automated, new requirements are emerging at the same time. Roles such as content architects or metadata specialists are becoming more important, as content must increasingly be approached systematically rather than purely editorially.

Another aspect became particularly clear: fewer and fewer users are accessing websites directly at all. Content is delivered via search engines, AI overviews or other systems and often appears outside its original context. As a result, it is no longer sufficient to optimise content solely for one’s own website. Today, content must also function outside one’s own platform whilst remaining comprehensible.

Visibility means more than just being found

In the past, the main focus was on appearing as high up as possible on a search results page. However, the way people find information is now changing significantly, and this was precisely the focus of the session "How AI is Reshaping Discovery, Websites, and Personalization" by Nicole Rogers, CMO and co-founder at ai12z.

Particularly interesting was the perspective on AI as an additional layer between the brand and the user. Today, content is not merely displayed, but interpreted, summarised and reclassified. This creates a situation that many brands have scarcely encountered before: they are increasingly losing full control over the context. Whilst content can be deliberately designed on a brand’s own website, AI systems decide for themselves which aspects to highlight and which information might be omitted.

The idea of not only clearly defining what a brand stands for, but also what it deliberately does not stand for, was therefore particularly valuable. Otherwise, when information is interpreted, connections can quickly arise that were never intended. Visibility alone does not automatically become the goal. Rather, what becomes crucial is how a brand is perceived and understood.

When visitor numbers drop, every single visitor becomes more important

One trend ran through many of the sessions at the CMS Summit and was also highlighted in the presentation by Søren, one of our managing directors: an increasing number of information journeys are coming to an end before users even visit a website. People are increasingly getting answers directly via AI overviews or generated search results, and as a result often no longer visit a website at all. This is fundamentally changing the role of the click.

People who visit a website today often have clearer expectations, search more specifically and, at the same time, have less patience. The focus is therefore shifting away from maximising visitor numbers and increasingly towards making the most of the visits we do receive. One tool that can help in this area is rek.ai, which can be integrated via the extension we have developed. Based on various factors, relevant content and recommendations can be displayed automatically, guiding visitors more effectively through the website.

The key insight here was deliberately kept independent of individual tools: websites should not only be optimised for AI, but above all for real people. After all, AI can change the path to the website, but the actual experience still takes place on the website itself.

Sounds exciting? Feel free to get in touch with us! 

More than just talks

All the sessions were of a very high standard. The line-up of speakers was impressive, the topics varied widely, and the content was technically in-depth. When it came to AI in particular, it was clear that the CMS Summit wasn’t simply about celebrating new tools. Instead, many discussions centred on responsibility, quality, critical scrutiny, and the question of how digital work can be designed in a meaningful way for the long term.

That is precisely what made the CMS Summit 2026 so special. The combination of curiosity, openness and a simultaneously reflective view of technological developments gave the event a particular depth. Instead of blindly chasing every trend, ideas were tested, discussed and, at times, challenged. And, to be honest, it is often precisely these conversations that are the most exciting.

Conclusion

For everyone working in the fields of content management, CMS platforms and digital communication, the two days were therefore not only inspiring, but above all practical and relevant to their day-to-day work. 

Special thanks go to Janus Boye and the entire team behind the event, which took place in Frankfurt for the second time. The CMS Summit 2026 has once again demonstrated that good digital work is created where expertise, exchange and diverse perspectives come together.