I’ve recently been living in two worlds. In one, which I find on LinkedIn, AI is praised to the skies and everyone has been an expert for a long time. In the other, an obscure nerd universe called Fediverse, the bubble is about to burst and AI slop is clogging up the internet.
It’s high time to take a closer look and shed light on the status quo of AI use in agencies towards the end of 2025. From a wide variety of conversations and research, we have learned that most agencies in the IPRN, and certainly outside it, are working on similar challenges and following a similar path of AI transformation. In large corporations and the network agencies that support them, there are numerous examples of advanced models and projects in which enormous resources are being devoted to AI implementation – but sometimes one gets the impression that even there, they are just cooking with water, as we say in Germany. Behind the buzzwords of abstractly described cases, it is not always possible to discern the real value and scope of the specific AI solution.
At TDUB, we started our approach quite early, but with a somewhat critical attitude. That’s why we initially placed great emphasis on a compliant use. It was only in the course of our work with AI that more and more application opportunities opened up to us, so that today we enthusiastically discuss new solutions almost every day.
What are the hot topics out there?
From our perspective, three main topics are currently shaping the discourse on the use of AI in agencies:
- Efficiency: How can the use of AI tools be made consistent and efficient? This involves moving from individual experiments with specific tools to consistent AI-supported PR work in everyday agency life.
- Innovation: How can we jointly develop valuable, new, AI-based offerings? This involves automating processes and exploiting all potential through specific tools such as agentic AI.
- Seizing opportunities: How can we secure our clients’ online presence in light of the rapid shift in power from search engines to LLMs? PR provides the perfect tools for generative reputation management and generative engine optimization (GEO).
For us at TDUB, AI has become an integral part of our daily work. We have trained our team specifically, introduced various tools, and discarded some of them after a trial period. We discuss the topic every single day. We continue to invest a lot of time in ensuring that our approach is legally and practically sound. The focus of daily use is certainly on research and brainstorming, as well as text production. It is clear that we currently achieve the greatest efficiency gains in routine tasks and short texts. The reformulation of texts in the communication process is extremely useful, for example when we create customized pitch papers for trade journalists, including the first draft of an email, directly from technical white papers provided by the customer. When creating longer or more complex content, the manual effort remains high—AI often only provides raw material that requires intensive post-processing. Too often, the central theme is lost, or the specific content is lost. Here, humans with their evaluation skills are simply irreplaceable. However, final checking, editing, and curating by “HI” remains a must anyway.
Agencies can develop new tools and offerings
We are seeing the possibilities evolve almost daily. The next logical step is to automate the processes. As is well known, the fight against hallucinations has led to tools such as ChatGPT and Perplexity now also evaluating and verifying the most up-to-date sources. This enables them to be used for agenda surfing. Our team is currently developing a process that compares and evaluates the media agenda with the client’s messages in real time, so that within minutes we receive tailor-made quotes that we can send to journalists to supplement their current reporting. In this context, we are testing agentic AI tools that can automate certain tasks, such as identifying agenda surfing opportunities or evaluating PR measures in relation to AI search results. The breathtaking range of possibilities that AI offers us here only becomes apparent as we delve deeper and deeper into the subject. However, developing such solutions is complex and requires expertise that we are building up in a targeted manner and may soon purchase. Ultimately, this could result in an AI-based service that we offer to our customers as an additional service. We firmly believe that agencies should be compensated for the expertise they build up.
AI search: GEO as a new opportunity for positioning
Meanwhile, there is a growing awareness that PR agencies are the ideal partners for “Generative Engine Optimization” (GEO) because they have mastered the core of earned media. This is because LLMs place a particularly high weight on trustworthy, organic content, as it is considered credible and relevant. PR agencies place brand messages in journalistic articles, trade media, and influencer posts – precisely the sources that LLMs prefer to include in their training and response logic. In this way, they increase the visibility and authority of a brand in “generative search” without relying on paid advertising. For us, this is an exciting field that we are increasingly entering.
An important point when working with clients: we work transparently and within clear compliance rules. For sensitive tasks, we only use premium models in which no client data is incorporated into the AI training. The introduction and subsequent revision of binding AI operating instructions for our employees was essential for this.
A mandate for cooperation between client and agency
We see a major task ahead: customers and service providers must come together and jointly define the path to AI transformation. As was the case with the establishment of new media and networks, agencies are taking on the role of researchers and mediators: they will increasingly advise their customers on the use of AI, share results and experiences, and develop new ways of working together.
This will enable agencies to leverage their new skills to add value and continue to be relevant to their clients’ communication success in the future. They are no longer just consultants in communication matters, but are once again supporting the efficient use of new technologies. They are not only content curators and irreplaceable relationship managers, but also process optimizers and “AI whisperers.” Agencies are not becoming cheaper, and they are certainly not becoming superfluous. Thanks to the skills they have gained, they are becoming increasingly valuable to their clients.
Tilo Timmermann, Founder & Managing Director at TDUB Kommunikationsberatung