evidoria

← Back to browse

Good practice Imported

Hugo.lv — Latvia's National Network of AI Virtual Assistants for Public Services

Latvia · Riga · See the Latvia profile · See the Riga profile

Evidence: Descriptive / self-reported Top 22% 73/100 · Ask Evidence Copilot about this practice

Latvia's Culture Information Systems Centre and vendor Tilde run national government chatbots — Signe, UNA, Ieva, Toms and others — spanning 40 public bodies. Signe alone handles over 25% of one agency's inquiries; UNA auto-resolves 44% of registration questions.

more than 25 %
Signe share of LVRTC customer-support inquiries handled
44 %
UNA questions auto-resolved without human handling
more than 10
Virtual agents in the network (by 2020)
over 2,000
Customers served per day across the network (by 2020)
40
Public bodies covered by the network (by 2020)
more than 6,100
Subjects covered across all assistants (by 2020)
3-4x
Usage increase during COVID-19
Hugo.lv — Latvia's National Network of AI Virtual Assistants for Public Services

Details

Maturity
Established
Promoter
Culture Information Systems Centre (VISC) & Tilde
Period
2018–present
Region (NUTS)
LV006
Keywords
digital government, citizen services, conversational AI, tax administration

Context

Since 2018, Latvia has built a national network of government virtual assistants on the Hugo.lv language-technology platform, developed by Latvian company Tilde with the Culture Information Systems Centre and Microsoft Latvia, and funded with European Union support.

Activities

Rather than one generic chatbot, each public body gets its own named assistant: UNA at the Register of Enterprises answers business-registration questions and lets users track applications, Signe at the Latvian State Radio and Television Centre handles general customer support, Ieva at the Rural Support Service explains complex agricultural-subsidy rules, and Toms at the State Revenue Service supports tax queries and supplies agency staff with analytics on common questions.

Results

By 2020 the network had grown to more than 10 virtual agents serving over 2,000 customers a day across 40 public bodies and covering more than 6,100 subjects; Signe handles more than 25% of all incoming customer-support inquiries at LVRTC, with usage rising 3-4x during COVID-19, and UNA auto-resolves an estimated 44% of the questions it receives.

Conclusions

These are largely vendor- and platform-operator-reported figures rather than independent third-party evaluations, and public detail on error rates, language coverage beyond Latvian, or citizen-satisfaction surveys remains limited.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Long (> 3 years)
Staffing & skills
Tilde (vendor and language-technology developer), Culture Information Systems Centre (VISC), Microsoft Latvia (technical partner)

Conditions for success

  • A centralised national language-technology platform (Hugo.lv) is reused across agencies rather than each public body building a bespoke chatbot.
  • Sustained European Union funding support for the platform since 2018.
  • Each public body gets a dedicated named assistant tailored to its own domain, such as business registration, tax, agriculture or broadcasting.

Common failure modes

  • Independent evaluation of accuracy, error rates and citizen satisfaction has not been published in available public sources.

Commonly funded by

National / regional programmes

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