AI for the local public sector

AI for municipalities in Switzerland

Municipalities and local governments: we deploy AI to better serve citizens and lighten the administrative load, with the data-sovereignty standard the public sector demands.

Talk to an expertDescribe my needs
European hostingOpen source modelsAuditable

14'036+ sites created in the last 30 days

actif
🇪🇺 Europe
Falkenstein
Helsinki
Nürnberg

Hetzner · Europe

US cloud
European hosting
Open source models
Auditable

Swiss municipalities face a difficult equation: growing administrative workload, a shortage of qualified staff, and rising citizen expectations, all within constrained budgets. Artificial intelligence is not a distant promise; it is an operational lever that hundreds of European administrations are already deploying to reduce processing times, free up staff from low-value tasks, and provide a service accessible around the clock. Kleap supports Swiss municipalities through this transition with a gradual, sovereign approach centred on the public servant.

14'036+
sites created in the last 30 days
16
languages
100%
European hosting
0
US cloud

AI in the service of citizens

A faster public service, with data under control.

Citizen services

Information, procedures, answers to common questions: AI helps your residents while your staff stay in control.

Lighter admin

Request processing, correspondence, document management: we automate repetitive tasks.

Sovereign data

Open source models on European infrastructure: citizens' data does not go to the US.

Transparency

Traceability of automated processing, for a responsible public service.

Why AI is becoming essential for Swiss municipalities

Pressure on municipal administrations is intensifying: backlogs in processing requests, recruitment difficulties, an explosion in document volumes, and citizens accustomed to instant digital services. The Swiss Association of Municipalities explicitly acknowledges this: AI will enter municipal administration, and it is better to anticipate it than to be caught off guard. The Federal Council has adopted a national AI strategy for public administration and strengthened the Competence Network for Artificial Intelligence (CNAI). The Swiss Digital Administration (ANS), founded in 2022, coordinates digitalisation at all levels, including the Confederation, cantons, and municipalities. Pioneer cantons such as Vaud and Genève have published practical guides. In Solothurn, Nidwalden, and Obwalden, algorithms are already partially processing tax declarations. The movement has begun.

  • Shortage of qualified administrative staff: AI compensates without replacing
  • Growing document volumes: forms, letters, minutes, certificates
  • Citizens demanding faster response times and more available services
  • Federal national AI strategy 2024-2027: framework and resources for municipalities
  • Cantonal guides available (Vaud, Genève): adoption is now possible and well-mapped

Concrete use cases by functional area

AI in a municipality does not mean automating everything at once. The fastest gains come from high-volume repetitive tasks. Here are the most mature applications, organised by area, that Kleap can deploy or support.

  • Citizen reception and service: AI assistant answering questions about regulations, procedures, deadlines, and forms, available outside office hours
  • Document management: automatic sorting, filing, and data extraction from scanned paper forms or incoming correspondence
  • Letters and correspondence: automated drafting of standard letters (responses, certificates, notices, reminders) with human validation before sending
  • Minutes: automatic generation of council meeting minutes from an audio recording
  • Finance and budget: automatic analysis of MCH2 financial data, budget variance alerts, dashboards
  • Human resources: drafting job descriptions, managing work certificates, searching collective agreements
  • Translation and accessibility: translation of official documents into several national languages, simplification of administrative language
  • Urban planning and one-stop shop: pre-screening of dossiers, completeness checks for permit applications

Data sovereignty and LPD compliance: the non-negotiable requirements of public service

Swiss public services operate under strict compliance constraints. The Data Protection Act (nLPD), in force since 1 September 2023, sets specific requirements for the processing of citizens' personal data. An AI solution deployed in a municipality must meet four fundamental criteria.

  • European hosting: infrastructure on servers located within the European Union or in Switzerland, with no transfer to American clouds (AWS, Azure, Google Cloud)
  • Open-source models: use of AI models whose workings can be audited and which run on your own infrastructure, without sending data to third-party APIs
  • Traceability and audit: every AI action must be logged, timestamped, and linked to the staff member who validated it
  • Human validation: no sensitive action (sending an official letter, an administrative decision) without staff supervision
  • Kleap uses infrastructure hosted by Hetzner (Falkenstein, Nuremberg, Helsinki, European Union), open-source models executed locally, and an architecture that involves no data transfer to third-party providers.

The real challenges of AI in the public sector and how to anticipate them

Municipalities that successfully deploy AI share a common approach: they start by identifying real pain points before choosing a tool. Here are the most frequent obstacles and concrete responses.

  • Reliability of responses: AI can produce errors or imprecise information. Solution: systematic supervision by a staff member, no official response without human validation
  • Resistance to change: some staff members fear for their jobs. Solution: involve personnel from the diagnostic phase; show that AI takes on tedious tasks, not responsibilities
  • Internal digital divide: uneven skills among staff members. Solution: training by profile, practical workshops based on real cases from the municipality
  • Algorithmic bias: risk of unequal treatment of citizens. Solution: regular model testing on representative datasets, human supervision for sensitive decisions
  • Cost and budget: smaller municipalities fear an unaffordable solution. Solution: start with a single use case at predictable cost, then expand progressively

Ethical governance and usage policy: the foundations of responsible deployment

Before activating any AI tool, a municipality must define its governance framework. This is not a barrier but a condition for lasting success. The essential elements of an AI policy for public services include: defining permitted and prohibited uses, human validation rules by type of decision, procedures for reporting errors or bias, minimum training required before accessing the tools, and the modalities for annual audit and review. This framework protects citizens, staff, and the municipality alike in the event of an inspection or dispute. Kleap can assist in drafting this policy as part of the deployment phase.

  • Clearly delineate: which processes can be automated and which remain 100% human
  • Define supervision levels by risk (low, medium, high)
  • Train and authorise: only trained staff members can access AI tools
  • Audit regularly: quarterly review of logs and AI-assisted decisions
  • Communicate with citizens: transparency about the use of AI in processing their requests

Our four-phase deployment process

A successful AI deployment in a municipality follows a logical progression, from diagnosis to steady-state operation. Kleap structures its support around four phases.

  • Phase 1 - Pain point diagnosis: audit of your existing processes, identification of the 3 to 5 most time-consuming and most automatable tasks
  • Phase 2 - Targeted pilot: deployment on a single use case, measurement of results over 4 to 8 weeks, adjustment
  • Phase 3 - Training and adoption: practical workshops by staff profile, drafting of the usage policy, implementation of validation procedures
  • Phase 4 - Roll-out and governance: gradual deployment on subsequent use cases, implementation of tracking dashboards, quarterly review

Three ways to deploy AI in your municipality with Kleap

Kleap offers three support models adapted to the size and resources of your administration.

  • Kleap builds for you: our partner agency Lionscreative takes charge of the full design, development, and deployment of your custom AI solution. Ideal for complex projects or groups of municipalities
  • Connection with a certified provider: if your municipality prefers to work with a local partner, Kleap connects you with qualified providers in French-speaking and German-speaking Switzerland
  • Kleap Enterprise self-serve: for IT managers and municipalities with internal resources, access to the Kleap platform to build and deploy your own AI tools, with support

Specificities of the Swiss context: what sets a Swiss municipality apart

Swiss municipalities operate within a particular legal, institutional, and linguistic framework that any AI solution must take into account.

  • Multilingualism: Switzerland has four national languages. A bilingual municipality (Fribourg, Berne, Valais) needs an AI capable of working in multiple languages without loss of quality
  • Municipal autonomy: each municipality defines its own processes. An AI solution must be flexible enough to adapt to local regulations
  • MCH2 standard: Swiss public accounting follows the MCH2 model. AI financial analysis tools must be familiar with this standard
  • LPD (nLPD 2023): the federal data protection act is stricter than the European GDPR on certain points. AI must be configured accordingly
  • Direct democracy: minutes of municipal councils, assemblies, and votes are subject to strict formal requirements
  • Militia system: many local elected officials hold their mandate part-time. AI tools must be accessible without heavy technical training

How an AI project with Kleap works in a municipality

01

Initial contact and scoping (week 1)

A discussion with your administrative manager or CIO to understand your current processes, your constraints (budget, timeline, LPD), and to identify the 3 to 5 priority use cases. Free of charge, no commitment.

02

Diagnosis and proposal (weeks 2-3)

A rapid audit of your existing processes. A costed proposal covering scope, timelines, training arrangements, and measurable success indicators.

03

Pilot on one use case (weeks 4-12)

Deployment on the fastest scope to validate (for example: citizen FAQ assistant or automation of standard letters). Measurement of results, adjustments.

04

Staff training and usage policy (weeks 10-12)

Practical workshops tailored by staff profile. Collaborative drafting of the AI usage policy. Implementation of validation and audit procedures.

05

Roll-out and follow-up (month 4+)

Deployment on subsequent use cases based on pilot results. Tracking dashboard. Quarterly review and evolution of the programme.

Kleap vs. generic solutions: what changes for a municipality

Not all AI solutions are equal when it comes to citizen data and regulatory compliance. Here are the criteria that matter.

CriterionKleapMass-market tools (ChatGPT, Copilot...)
Data hostingEuropean Union (Hetzner)USA (Microsoft, Google, OpenAI)
LPD/GDPR complianceArchitecture designed for public servicesMass-market tools, requiring adaptation
AI modelsOpen source, auditable, executed locallyProprietary models, black box
Traceability and auditFull logs, timestamps, human validationVariable, often limited
Customisation to local regulationsConfigured on your official documentsGeneric, no local knowledge
SupportAgency, training, usage policy, pilotNone (self-service)
Swiss multilingual supportFR, DE, IT, RM nativeVariable depending on the tool

Sovereignty

Citizens' data stays in Europe

The public sector cannot entrust citizens' data to just anyone.

European hosting

Infrastructure in Europe (Hetzner), no US cloud.

Open source models

Run on our infrastructure, not via a third-party API.

Auditable

Every operation is logged and verifiable.

swissIa.iaCommunesSuisse.localContextTitle

French-speaking Switzerland: Vaud, Genève, Fribourg, Valais, Neuchâtel, and Jura have published or are about to publish AI guides for their municipalities. The Association of Vaud Municipalities (ACV) produced an AI guide as early as 2024.
Swiss Digital Administration (ANS): the Confederation-cantons-municipalities coordination body is steering the 2024-2027 digital strategy, of which AI is an integral part.
AI Competence Network (CNAI): attached to the Federal Chancellery since 2026, it provides resources and recommendations to administrations at all levels.
Swiss AI Center (HES-SO): founded at the end of 2023, it specifically supports Swiss SMEs and administrations in adopting AI.
Concrete precedents: the municipalities of Solothurn, Nidwalden, and Obwalden are already using algorithms to partially process tax declarations.
Typical example: a municipality that implements an AI document management system can significantly reduce its mail processing time, with fully traceable processes.

Frequently asked questions

Is AI in a municipality compatible with the Swiss LPD?

Yes, provided the right architecture is chosen. The nLPD (in force since September 2023) requires, among other things, data minimisation, traceability of processing, and no transfers to countries without adequate protection. Kleap uses open-source models hosted in Europe (Hetzner, EU), without sending data to American APIs, and with complete audit logs. This architecture was designed to meet these requirements.

Where should a small municipality with limited IT resources start?

The best starting point is a single use case with high volume and low risk, for example a citizen FAQ assistant that answers questions about opening hours, forms, and routine procedures. The deployment does not require internal IT skills: Kleap or its partners handle the configuration. The cost is predictable and the pilot can be validated in 6 to 8 weeks.

Will AI replace municipal staff?

No. AI handles repetitive, low-value tasks: sorting mail, drafting standard letters, answering frequently asked questions. Staff members get back time for complex situations, citizen advisory work, and decisions that require human judgement. In pilot municipalities, staff describe AI as a tool that 'does the paperwork' so they can do their real job.

Does citizen data leave Switzerland or the EU?

No. Kleap's infrastructure is hosted by Hetzner, whose data centres are in Germany and Finland (European Union). The AI models used are open source and run on this infrastructure without sending data to third-party providers (OpenAI, Google, Microsoft). No citizen data passes through American servers.

What is the difference between Kleap and a tool like Microsoft Copilot?

Microsoft Copilot is a general-purpose tool designed for private enterprises. It processes data on Microsoft servers (primarily in the USA), with compliance guarantees that require complex additional configurations and specific licences. Kleap is architected from the outset for public services: EU hosting, auditable models, mandatory human validation, and configuration based on your local regulations and procedures.

How should resistance to change among staff be managed?

Experience from successful deployments shows that staff buy-in requires three conditions: being involved in the choice of use cases (they know better than anyone which tasks are burdensome), receiving practical training based on examples from their own municipality (not a generic demonstration), and quickly seeing that AI does not threaten their position but reduces pressure. Kleap systematically includes adoption workshops in its support programme.

Can a bilingual municipality use Kleap?

Yes. Kleap supports French, German, Italian, and Romansh. Citizen assistants can be configured to respond in the language chosen by the user. Municipalities in bilingual cantons (Fribourg, Berne, Valais) and trilingual cantons (Graubünden) are natural use cases for these features.

Is a significant IT budget needed to get started?

No. Kleap offers a gradual start on a limited scope, with predictable costs from the outset. The first pilot (one use case, 6 to 8 weeks) is sized for standard municipal budgets. Expansion happens at the pace of results and available resources.

How does Kleap ensure the reliability of AI responses?

Kleap applies two principles: first, the AI is trained on your own official documents (regulations, directives, internal FAQs), not on generic data that might be incorrect or outdated. Second, any sensitive response or action goes through human validation before being sent to the citizen or executed. AI errors are systematically logged for continuous improvement.

Can Kleap integrate with our existing municipal management software?

In most cases, yes. Kleap has connectors for the main solutions used in French-speaking Switzerland (document management systems, municipal ERPs, messaging). Specific integrations are assessed during the diagnostic phase. If your software exposes an API or standard exports, integration is generally achievable.

Related topics

AI for SMEs in SwitzerlandAI for nonprofits in SwitzerlandSovereign AI in Switzerland

A public service enhanced by AI

Let's talk about your municipality's needs. We propose a concrete, sovereign approach.

Request a Custom Demo

Tell us about your team and we'll reach out within 24 hours.

We'll never share your information. Expect a response within 24h.

AI for Municipalities in Switzerland | Local Public Sector