What does Neuraflow stand for?
Our approach is to make administrative expertise more transparent and accessible through semantic search technologies and large language models (LLMs) – for citizens and employees alike. So far, we have implemented three products for local authorities: an AI chatbot and an AI voicebot for citizen communication, as well as an internal AI assistant. We are working with our partner towns and cities on further applications, for example in knowledge management.
What are the towns and cities’ expectations of you?
Often, they have high expectations – especially when it comes to AI, which is new territory for many. Many see it as a huge wave that they feel overwhelmed and overrun by. We see it as our job to show them how to ride this wave. But even if decision-makers have recognized the added value, broad political and administrative support is needed. And ultimately, it all comes down to implementation: Younger employees are usually particularly quick to embrace change, while older colleagues often react more cautiously at first.
©neuraflow GmbH
Is commitment the biggest hurdle in implementing municipal projects?
No, on the contrary. Municipal employee commitment is the biggest driver of digitalization. The biggest challenges are clearly financing and procurement procedures, which can be very complex and time-consuming at the municipal level. This sometimes leads to curious situations where administrations circumvent complex procurement processes and award AI projects to framework agreement partners who do not have the required expertise and references. Fortunately, federal states such as Hamburg and Baden-Württemberg have now introduced regulations that make it easier to award projects worth up to €100,000 to start-ups. To make matters worse, innovation projects in many towns and cities can only be commissioned if cost-saving potential can be demonstrated.
What would need to happen for local authorities to be more open to innovation?
Above all, more courage would help. In companies and start-ups, an iterative approach – i.e., gradual testing, learning and adaptation – is a matter of course. In public administration, on the other hand, thinking in terms of large-scale projects with advance planning often dominates. It would be important for courageous steps to be rewarded there as well. Interestingly, we often find this courage in towns and cities with between 30,000 and 80,000 inhabitants. They usually have the necessary resources for digitalization and short decision-making processes. In large cities, on the other hand, decisions often have to be coordinated across several hierarchical levels, which means it can take a long time for them to be implemented.
You bring a wealth of experience from large projects. How important is trust?
Trust is everything. We rely on very close cooperation with the digitalization teams of our partner towns and cities so that both sides benefit. We usually offer a non-binding trial period to towns and cities that don’t know us well yet. We also always focus on transparency: This means that we document exactly where our AI solutions get their answers from – every statement is linked to a source. This allows administrations to understand the information on which the answers are based.
Apart from financing and trust, where are the sticking points in project implementation?
We encounter problems in implementation where the quality of existing data is poor. Typical examples include missing, outdated or contradictory information on a municipal website or in other municipal sources. AI cannot compensate for missing content – nor do we want it to. That’s why we work closely with the specialist departments, because they know the quality of their data best.
Data protection is a sensitive issue for local authorities. How do you deal with it?
When communicating with citizens, we work exclusively with publicly available data from the local authorities and take great care to ensure that our chatbots and voicebots do not process any sensitive personal information. In addition, we host our AI solutions exclusively on certified German servers with a German cloud hosting provider. Local authorities can also choose whether they want to use market-leading LLM or open-source alternatives, depending on their requirements in terms of data protection, costs and performance. This allows them to choose the degree of their digital sovereignty themselves.
We talked about obstacles. What does it take to make projects successful for both sides – local authorities and providers?
It is crucial that local authorities do not just wait for funding measures, but invest themselves too: Digitization measures have long-term added value for citizens and also offer opportunities for savings. The use of generative AI in public administration can simplify communication and information work and reduce the workload on teams. This gives employees more time for complex consultations and tasks where human expertise and empathy are irreplaceable.
AI can streamline processes – what would fundamentally help administrations to work more efficiently and sustainably?
It is essential to involve employees at all levels from the outset – not just decision-makers. If employees are left out, there will be a lack of acceptance. Here is an example: In one city, the digitization of files was accelerated by having employees’ older children earn some extra money by helping out with scanning. They talked about it at home and thus brought their parents into the digitization process.
Looking five to ten years ahead, where do you see Neuraflow?
We want to be present in all areas of administration where AI solutions can create added value. Our vision is an intelligent AI platform that networks municipal knowledge and makes it usable both internally and externally.
Do you think Germany is capable of making this leap in the next ten years?
Germany faces particular challenges – the federal system requires more coordination than is the case in centrally organized countries. But there are currently many political initiatives at the federal and state levels to make things leaner, more effective, and more uniform. It remains to be seen which of these will have an impact at the local authority level. But we are working precisely on the approach that is repeatedly mentioned in the Federal Ministry for Digital Transformation and Government Modernisation: identifying successful solutions, scaling them up, and making them available to other local authorities. And we see every day that there are innovative cities and towns that are showing what is possible.
Neuraflow was founded in Bremerhaven in 2024 and develops AI-supported solutions such as chatbots, voicebots and smart search systems for public administration. The aim is to advance digitization in the public sector and simplify communication with citizens.