Expert Interview: AI as a game changer in public services

Key Theses

• AI has already arrived in public administration – according to Hanl, the focus is no longer on "if", but on "how"
• AI applications such as chatbots or automated tender procedures can ease the burden on authorities and improve citizen services
• Fears and uncertainties about AI often arise from a lack of strategy – concrete pilot projects can help build trust
• Hanl sees the future of administration in a citizen-centered digital transformation – AI is a key building block in this

About Moritz Hanl

Moritz Hanl is an expert in implementing AI chatbots in the public sector and large corporations. After studying business administration and gaining experience as a tech founder, he specialized at viind in introducing tailored, GDPR-compliant AI solutions. Since 2023, he has supported municipalities and companies in optimizing their digital communication, implementing various successful AI projects.

Mr. Hanl, is German administration even ready for AI in 2025, or is that still wishful thinking?

German administration has understood the potential of artificial intelligence from day one. Since then, a wide range of use cases have been implemented. Our AI chatbot solution alone is being used in over 30 major cities, counties, and small municipalities. In our conversations with the public sector, it has never been about an “if”, but always about a motivated “how”. The increasing adoption of AI in authorities is also evident from other examples: For instance, GovRadar, which helps create tenders using AI. SummAI, which converts texts into plain language. And the city of Leipzig, which developed a housing benefit form with LECOS using AI.

Can you give examples where AI has already made a noticeable difference in a German authority?

I want to emphasize that my specialization is in AI chatbots. A prime example is our client, the district of Regensburg: Since making a chatbot accessible via their website and WhatsApp, as well as a telephone-based AI bot, the number of inquiries requiring human response has dropped significantly. We built a chatbot that connects a specific data base with large language models like those used by GPT. This allows the bot to be integrated into a city’s website and answer citizens’ questions about the city. The bot only accesses content that has been pre-approved. In a second step, the software can be deployed across different channels: WhatsApp, phone, and the website. And all of it is GDPR-compliant. You can see it in action on the websites of Augsburg, Memmingen, or the district of Regensburg.

What holds authorities back when it comes to implementing AI?

The main issues have remained the same for years: tangled bureaucracy, inefficient silo solutions, and lack of investment resources. Specifically for AI, there's an added hesitation due to the absence of clear AI strategies. This hesitation mainly shows up as caution: there's fear of AI making mistakes — for instance, giving wrong answers. People naturally fear the unknown, and AI is very new. That fear is understandable, and at the same time, we know how to tackle fear: through exposure. That’s why I’m thrilled that more and more representatives in the public sector are testing AI in various areas. Only this way can we create a better data foundation for expectation management together.

What are your tips? Where should resistance and fear be taken seriously, and where can they be easily dispelled?

Of course, we need brave pioneers in the field of AI. At the same time, I fully understand that this role is hard to take on, especially due to limited resources. The good news is: there are more and more reference projects. Start communicating, share experiences, and take advantage of synergy effects. If resources are especially tight, choose projects with a clear return on investment.

How can we successfully implement solutions from start-ups and established software providers in government agencies?

Authorities are very cautious and thorough. That’s a good thing. If a solution works well in one use case, it often adds value in other agencies as well. That’s why I refer back to my previous answer and advocate for more exchange. In addition to existing formats, nationwide marketplaces of tested software solutions would be very exciting for this.

When used properly, what tasks can AI take off the hands of authorities? And where does it reach its limits?

AI offers a completely new world of process optimization. Suddenly, there's no such thing as “too much” data or content — in fact, AI thrives on it, unlike humans. I hope we can use this opportunity not just to digitize processes, but to rethink them entirely. The limits of AI are clearly in the human realm. Traits like human empathy, decision-making authority, and physical presence will never be replaced by AI.

Many citizens experience public administration as complicated and slow. Can AI change that – or is bureaucracy just bureaucracy?

I agree with your statement that many citizens perceive administrative processes as complicated and slow. The real question is “why?” As long as that remains unanswered, AI can only alleviate the symptoms. The “whys” show up in federalism, accessibility, and bureaucracy. For every law, there’s another law – ideally one that varies by federal state. As a citizen, I have to know whether to go to the city or the district office, whether the process is available digitally, and which conditions I have to meet. That’s not a great user experience, as we would say in the private sector. But since these “whys” are mostly identified by now, I’m confident we’ll revolutionize the citizen experience in dealings with authorities in the coming years.

Imagine it's the year 2035: What role does AI play in public administration? What developments would be desirable – and which ones maybe not?

Now I feel like I’m in a job interview. I like the thought experiment. For me, it’s not about how much AI is in the processes — the term is broadly defined and often open to interpretation anyway. What’s more important is that the public sector sees itself as “Governance as a Service” by then. Citizens should have to worry as little as possible and enjoy a great user experience when interacting. My utopia is that citizens actually look forward to (digital) interactions with authorities. Why not?

If you could digitize a ministry or a city tomorrow – which three AI projects would you implement immediately?

Haha, totally unbiased, I would recommend internal and external AI chatbots, ideally from viind. Aside from that, I follow developments in digital twins, identity verification, and data analysis with great interest. Every institution knows its own challenges best. I’ll go out on a limb and say: there’s already an AI solution on the market for your problem. So let’s optimize and test them together!