evidoria

← Back to browse

Good practice

ChatbotPJ — Costa Rica's Judicial-Branch AI Assistant for Case and Service Queries

Costa Rica · San José · See the Costa Rica profile

Costa Rica's Poder Judicial has run ChatbotPJ since 2018 to answer routine questions and, since 2023, judicial-collection case status 24/7. It reports about 5,500 queries a month and over $14,000 in monthly savings versus its call centre, plus a public AI-use disclosure.

5,500 queries/month
Frequent queries handled
>$14,000 USD/month
Monthly cost savings
916,000
Document classifier filings processed (2024)
$214,000 USD/year
Document classifier estimated savings
60+
Budget centres using ML budget-forecasting model
ChatbotPJ — Costa Rica's Judicial-Branch AI Assistant for Case and Service Queries

Details

Maturity
Established
Promoter
Poder Judicial de Costa Rica (Judicial Branch of Costa Rica)
Period
Launched 2018; case-status lookup added January 2023; ongoing as of mid-2026
Keywords
e-justice, citizen services, natural language processing, judicial administration

Context

ChatbotPJ, published at chatbotpj.azurewebsites.net and linked from the main Poder Judicial website, is an AI assistant for Costa Rica's Judicial Branch, created in 2018 to relieve congestion on the judiciary's 800-800-3000 helpline.

Objectives

Answer natural-language questions about court offices and procedures, and (since a January 2023 expansion) let users look up Cobro Judicial case status by case-file number around the clock.

Activities

Operates as one of several AI tools the Poder Judicial has deployed, alongside a document classifier and a 2019 machine-learning budget-forecasting model used by more than 60 budget centres. The Judicial Branch publishes a formal 'Declaración de uso de Inteligencia Artificial' on its transparency portal.

Results

Handles roughly 5,500 frequent queries a month, saving over $14,000 monthly per El Financiero; the related document classifier processed 916,000 filings in 2024 for an estimated $214,000 in annual savings.

Conclusions

Reporting covers usage volume and cost savings but not resolution accuracy or citizen satisfaction, which remain undisclosed; the pattern points to a sustained, multi-year internal AI programme rather than a one-off pilot.

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Data sources

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

Attachments

Similar practices you may find useful