Where public bodies get value first
Internal use cases with a clear document base work best: questions about administrative regulations and internal instructions, guidance on forms and responsibilities, summaries of long circulars, or finding the right template for a decision letter. The assistant proposes an answer with evidence, and the case worker decides what to do with it.
Citizen-facing chatbots are possible, but they raise the bar on transparency, accessibility and quality assurance. Many administrations therefore start internally and open up later.
Onboarding is another benefit. New hires and career changers get to grips with responsibilities, procedures and terminology faster, without constantly pulling experienced colleagues away from their work.
GDPR, the EU AI Act and information security
Public bodies in the EU must comply with the GDPR and national or regional data protection law. The EU AI Act requires transparency when people interact directly with an AI system and classifies some systems as high risk, for example those used to decide on access to public services. A staff knowledge assistant that makes no decisions is much easier to assess.
We document data flows, roles and permissions so your data protection officer, security team and staff council can review the solution. Our processes are certified to ISO 9001 and ISO 27001.
Hosting in Germany or the EU and alignment with national security baselines, such as the BSI IT-Grundschutz in Germany, are common requirements. We settle these points before the pilot begins, together with a short AI literacy briefing for users.
From first idea to pilot
If needed, we also help you write the specification for a public tender, covering hosting, logging and role concepts, so that bids can be compared fairly.
- In a free 15-minute intro call we discuss the use case, users and constraints such as hosting and procurement.
- You select a well-defined document set, for example the internal guidance of one department.
- Our AI code engineers configure kalAIDoCreate and test it with real questions from case workers.
- Together we review answer quality, adoption and effort, then decide whether to scale.
Frequently asked questions
Does the AI make administrative decisions?
No. It supports staff with information and references. Decisions on applications remain with the responsible official.
Is this allowed under the EU AI Act?
An internal knowledge assistant without a decision-making role is generally not a high-risk system. Transparency duties and the obligation to ensure AI literacy among users still apply. We help you assess this, but we do not replace legal advice.
Do you work on site?
Yes. Our teams work remotely across the DACH region and internationally, and on site when it helps, for example in workshops with departments and the staff council.