How AI shows what really works in communication

Edzard Bennmann Head of Corporate Communications at SIGNAL IDUNA Group

Reputation analysis rethought: In our latest interview, Edzard Bennmann, Head of Corporate Communications at the Signal Iduna Group, explains how AI evaluates millions of posts, makes impact visible and helps communications departments to manage more strategically. An insight from the field – with clear added value for anyone who wants to further develop data-based communication.

“AI shows us what really works”

AG CommTech talks to Edzard Bennmann about how artificial intelligence is changing media analysis – and why it has long been more than just a reporting tool for communications departments.

AG CommTech: Edzard, you have been working with AI in media analysis for years. Why do we need it today?

Edzard Bennmann: Because traditional media monitoring has its limits. If you as a company want to know how you are really perceived, you have to understand millions of posts from earned, owned and social media – and not just that something is being said, but what it means. AI helps us to recognize correlations that no human could grasp with such speed and precision. We can immediately see which topics have an impact on our reputation, which risks arise and which measures are effective. This makes communication controllable instead of reactive.

AG CommTech: You work with a five-dimension model for reputation. What are the practical benefits of this?

Edzard Bennmann: Reputation is often discussed in abstract terms. The model – Employer, Board/Management, Products & Services, Profitability and Sustainability – makes it measurable in concrete terms. The AI assigns each individual statement in the network to these dimensions. This allows you to immediately understand where you are doing well, where you are losing trust and which factors are behind this. What is particularly valuable is that we recognize developments quickly: Ten years ago, employer branding was not a big issue for us. Today, it is one of the strongest drivers of overall perception. Without AI, we would have noticed this shift much later.

AG CommTech: How reliably does AI recognize such dimensions?

Edzard Bennmann: Very reliable – if it is properly trained. The analysis runs at fragment level, i.e. not on the entire article, but on the specific statement. The neural network first recognizes whether our company is actually meant – you wouldn’t believe how often brand terms can mean something completely different – and then precisely assigns which topic is affected and which tonality is present. This creates a realistic picture of perception, not just a superficial sentiment analysis.

AG CommTech: You can even use AI to track the impact of your communication work. How does that work?

Edzard Bennmann: The particular strength lies in the linking of all impact levels – input, output, outtakes, outcomes and ultimately reputation. We can see exactly what we have published, how visibility has changed, what response there has been and whether all of this has ultimately influenced our reputation. This turns communication work into a controllable process. The AI recognizes patterns, identifies promising communication channels and helps us to continuously optimize our measures. The result is a systematic look ahead, not just in the rear-view mirror.

AG CommTech: Can you give us a practical example?

Edzard Bennmann: A very vivid example is our change of CEO this year. We wanted to understand exactly how this transition would be perceived externally. The AI immediately showed us that the topic was one of the strongest reputation drivers. We were able to see which media were particularly relevant, how the waves of interviews were published and what tone prevailed.

In addition, we deliberately excluded our sponsorship of the stadium name – Signal Iduna Park – from the analysis right from the start. This topic is present in so many forums that it would have overshadowed everything else. The tool allows such adjustments to be made so that results are not distorted.

AG CommTech: Many people are worried that AI is hallucinating. How do you make sure that doesn’t happen?

Edzard Bennmann: The key is that we work on a verified data lake. The AI assistant only accesses verified, real data – earned, owned and paid data, including benchmarks. If it doesn’t know something, it says so. This prevents content from being invented, as can happen with open language models. This is hugely important for regulated industries such as insurance. AI helps us to work faster and more precisely without losing control mechanisms.

AG CommTech: How do you use the AI assistant in your day-to-day work?

Edzard Bennmann: Very pragmatic. I can ask very specific questions such as “Which media have had the biggest impact on our reputation?” or “Why has the sustainability dimension changed?” The AI immediately searches the entire database, providing sources, reach, tonality and correlations. This saves an enormous amount of time and enables a much more analytical approach. However, despite all the efficiency, human judgment remains crucial. AI is an aid, not a substitute for experience. An example: an in-house AI once claimed that we had long been CO₂-neutral – which was not true. So you have to stay awake.

AG CommTech: What does this development mean for the communications sector as a whole?

Edzard Bennmann: That the standard is changing. Reporting is no longer enough. AI is about making effects visible, recognizing developments at an early stage and managing communication with foresight. Teams work faster, more fact-based and more strategically. And they have better arguments – vis-à-vis the board, specialist departments and budget managers. If you can prove how communication contributes to reputation, employer branding or product perception, you will be given a different role in the company. AI helps us to take this step – and it is becoming increasingly important.

AG CommTech: Your most important learning for practitioners?

Edzard Bennmann: Don’t start with the technology, but with the question of what you want to control. Once this is clear, AI can unfold its full potential. It helps us to better understand what is happening outside – and to influence what happens tomorrow in a more targeted way. This is precisely where the big leverage lies.