Two Messages, One Answer: How a Prague Startup Is Testing AI on the Smallest Units of Czech Government

From Prague 6 to a village of a few hundred

Citymind, a Prague-based startup co-founded by Martin Mikulášek, builds AI chatbots for Czech municipal websites — virtual agents that answer citizen questions about office hours, permits, local events and traffic changes without routing every query to an overstretched municipal clerk. The company's highest-profile deployment sits in Prague 6, a district of roughly 100,000 residents, where a chatbot went live in June 2025 as part of CommuniCity, an EU-funded project coordinated with Finnish partners and delivered locally with the municipal company Operátor ICT. But Citymind's first Czech deployment predates that by more than a year: on 1 May 2024, the company launched what it describes as the first AI chatbot built specifically for citizen-facing service anywhere in the Czech Republic, in the small South Moravian municipality of Mokrá-Horákov.

The pairing is deliberate. A district-level deployment in Prague demonstrates the technology can handle the volume and complexity of a mid-sized urban authority; a village-level deployment demonstrates that the same platform is cheap and simple enough for a municipal office with a handful of staff to run without dedicated IT support. Citymind's own framing is that municipalities of any size can be digitally capable, not just the ones with in-house digital transformation budgets.

The numbers behind the pitch

Co-creation testing sessions run with Prague 6 residents produced average citizen ratings of 4.05 out of 5 on willingness to keep using the chatbot as-is, and 3.80 out of 5 on time saved compared with standard channels — scores Citymind contrasts with a national benchmark of 3.2 out of 5 that Czech citizens give the state's digital services generally, based on data from the civic-tech initiative cesko.digital. Separately, Ministry of Interior data on municipal responsiveness gives some sense of why the baseline is low: only 27 percent of municipalities respond to an enquiry within a week, and just 11 percent treat responding within the legal deadline — typically 15 to 30 days — as the ideal outcome rather than the ceiling.

Set against that baseline, an AI agent that resolves a typical citizen query in two exchanged messages is a meaningfully different service model, not simply a faster version of the same one.

What this adds to the national AI picture

The Czech government's NAIS 2030 strategy and its CZK 19 billion action plan set out public-administration AI as a national priority, but the concrete public-facing deployments so far — Citymind's municipal rollout among them — have mostly happened through EU project funding (CommuniCity), municipal IT budgets, and direct municipality-to-vendor contracts, rather than through a centrally coordinated NAIS rollout. That is not necessarily a criticism of the national strategy; procurement at the level of an individual town hall was never going to run through a national programme office. But it does mean the most-cited success metrics for AI in Czech public administration to date come from a single vendor's case studies rather than from independent government evaluation, and Citymind is candid that the market now includes a large number of competing companies offering broadly similar assistants to public bodies.

The scaling question

Citymind says it has ambitions to expand the same model beyond the Czech Republic and Slovakia, treating the domestic municipal rollout as a proof of concept rather than an end state. Whether Czech public administration itself scales the approach — moving from opportunistic, project-funded pilots in individual towns to something closer to a standard municipal software category — will depend less on the technology, which appears to work at both Prague-district and single-village scale, than on whether procurement and funding mechanisms at the Ministry of Interior catch up with what individual municipalities have already proven out on their own.