evidoria

← Back to browse

Good practice Imported

Eva and Jenda — the Czech Labour Ministry's AI Call-Centre Assistant and Benefits Portal

Czechia · Prague · See the Czechia profile · See the Prague profile

Evidence: Descriptive / self-reported Top 46% 60/100 · Ask Evidence Copilot about this practice

The Czech Ministry of Labour and Social Affairs built Eva, a generative-AI assistant for its benefits call centre, and added AI document-reading to its Jenda portal. The ministry and Microsoft report strong usage gains; no independent audit has yet verified the figures.

98%
Call-centre availability
55%
Inquiries resolved without human agent
60% → 90%
Benefit payments completed within 30 days (2021 vs. today)
100,000+
Clients served monthly via Jenda
Eva and Jenda — the Czech Labour Ministry's AI Call-Centre Assistant and Benefits Portal

Details

Maturity
Scaling
Promoter
Ministry of Labour and Social Affairs (MPSV)
Period
2023-2025
Keywords
social welfare, public administration, citizen services

Context

The Czech Ministry of Labour and Social Affairs (MPSV) runs Jenda, a self-service client portal launched in 2023 for benefits applications, alongside Eva, a generative-AI virtual assistant built on Azure OpenAI for its Call Centre for Social Benefits and Child Groups.

Activities

Eva greets callers and resolves routine questions, passing calls to a human operator when needed; Azure AI Document Intelligence automatically extracts data from paper application forms on the Jenda side, replacing manual keying by ministry staff.

Results

MPSV's lead IT architect said Eva lifted call-centre availability to 98% and cut operator workload by 40%; Microsoft's case study separately reports Eva resolves 55% of inquiries without a human and has cut call volume reaching agents by 60%. Microsoft also reports Jenda now serves more than 100,000 clients a month, with the share of benefit payments completed within 30 days rising from about 60% in 2021 to 90% today.

Conclusions

The quantified results published so far come mainly from Microsoft's customer case study and MPSV's own award submission rather than an independent government audit; MPSV's own website still describes Eva as 'in testing', and no third-party evaluation of accuracy, error rates or equity impact was found.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
MPSV IT architecture team, Microsoft (Azure OpenAI, Document Intelligence)

Conditions for success

  • Human fallback for calls Eva cannot resolve
  • Azure AI Document Intelligence for paper-form digitisation

Common failure modes

  • No independent audit of accuracy, error rates or equity impact; MPSV's own site still describes Eva as 'in testing'

Commonly funded by

Digital Europe Programme ESF+ — European Social Fund Plus

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful