Back to catalog
Dataset · Public services

Multilingual Public Services Inquiries

Synthetic citizen inquiries to public services across common topics — for testing multilingual AI assistants, intent detection and routing logic.

Verified for defined useDossier includedPublic services
multilingual-public-services-inquiries · v1.0 · signed manifest
Request access
FormatJSONL
LanguageMultilingual
Records6,500
Versionv1.0

Intended use

  • Testing
  • Validation
  • Demonstration

Not intended for

  • Automated decisions about real citizens
  • Training without independent review

Documentation level

High

Overview

Synthetic citizen inquiries in English and Serbian covering documents, taxes, benefits, healthcare and municipal services. Each inquiry is labelled with intent, expected routing and risk class, suitable for testing multilingual public-sector AI assistants.

What it contains

  • Dataset file (JSONL)
  • Sample preview (200 records)
  • Data dictionary
  • waiser data Dossier
  • License terms
  • Version log

Generation method

Generated via a scenario-based LLM pipeline with topic templates derived from publicly listed public-service categories. Each entry is reviewed for naturalness in both languages. No real citizen messages were used as seed.

Permitted and prohibited uses

Permitted

  • Testing multilingual AI assistants
  • Intent detection benchmarks
  • Routing-logic validation
  • Demonstration of public-sector AI pipelines

Prohibited

  • Automated decisions affecting real citizens
  • Training without independent quality review
  • Use outside the permitted license scope

Dataset structure

ColumnDescription
inquiry_idUnique synthetic inquiry identifier
languageen / sr
topicdocuments / taxes / benefits / healthcare / municipal
messageSynthetic inquiry text
intentLabelled user intent
expected_routeDepartment or workflow expected to handle the inquiry
risk_classLow / medium / high

Quality notes

Coverage targets the most common public-service topics. Less common edge topics are underrepresented and flagged in the dossier.

waiser data Readiness Profile

Documentation levelHigh
Permitted use clarityHigh
Privacy riskLow
Utility for defined useMedium
Ethical reviewCompleted

Ethical notes

Synthetic citizens do not represent real demographics. Do not use to make claims about service equity for real populations without additional evidence.

License

waiser data Standard License — Defined Use

Full license terms are provided with the dataset delivery. Access is granted after review of intended use.

Version log

v1.02026-05-20Initial release
Documented at every step

ready to request access?

Tell us your intended use and we'll respond with the executed licence, the full dossier and a secure delivery link bound to your request.