ARG-002

SRT 'Julieta' Chatbot for Occupational Risk Inquiries

Download PDF
Argentina Latin America & Caribbean Upper middle income Full Production Deployment Confirmed

Superintendencia de Riesgos del Trabajo (SRT)

At a Glance

What it does Perception and extraction from unstructured inputs — User communication and interaction
Who runs it Superintendencia de Riesgos del Trabajo (SRT)
Programme Sistema de Riesgos del Trabajo (Occupational Risk System)
Confidence Confirmed
Deployment Status Full Production Deployment
Key Risks Model-related risks
Key Outcomes Error margin reduced from 30% to 7% (93% accuracy) through focus group testing methodology.
Source Quality 6 sources — News article / media, Other, Government website / press release

Argentina's Superintendencia de Riesgos del Trabajo (SRT), the national public body responsible for regulating the occupational risk system, deployed an artificial intelligence chatbot named 'Julieta' in 2018 to handle citizen inquiries about work injury benefits and occupational risk procedures. Julieta was the first chatbot implemented in Argentina's national public administration and was recognised by the International Social Security Association (ISSA) as a Good Practice (gp/198017) in the Americas 2020 competition.

The chatbot was named in honour of Julieta Lanteri, a physician and women's rights pioneer who in 1911 became the first woman authorised by Argentina's Electoral Justice to vote in the municipality of Buenos Aires. The project was deliberately launched around International Women's Day (8 March), and the chatbot adopts a female persona depicted wearing a hard hat and safety glasses to reinforce workplace safety messaging. The conversational tone is colloquial and cordial, with the motto 'Prevention is everyone's daily work'.

Julieta was developed through a collaboration between the SRT and Microsoft Argentina, with the presentation taking place on 13 June 2018 at Microsoft's offices in Buenos Aires, attended by Microsoft Argentina General Manager Diego Bekerman and SRT Superintendent Gustavo Morón and General Manager Guillermo Arancibia. The chatbot is built on Microsoft's natural language processing technology, enabling it to understand human language and identify relevant information to answer user questions. The system features data validation mechanisms for sensitive information and requires identity verification before accessing personal worker records, connecting to the SRT's worker database to provide individualised responses.

The immediate catalyst for Julieta's deployment was the activation in February 2017 of Law No. 27,348, which reformed Argentina's occupational risk procedures and doubled the volume of consultations to the SRT's call centre. Faced with this surge in demand, the SRT opted for technological innovation rather than simply expanding call centre capacity. Julieta operates 24 hours a day, 365 days a year, accessible through the SRT website (www.srt.gob.ar) and social media channels.

Julieta addresses a comprehensive range of inquiries related to Argentina's occupational risk system. Users can check the status of their case files (expedientes), identify their assigned Aseguradora de Riesgos del Trabajo (ART, the occupational risk insurance provider), determine which medical commission they are assigned to, file complaints about delayed resolution of proceedings, and receive guidance on specific procedures including incapacity determinations, claim rejection appeals, benefit coverage divergences, treatment re-entry requests, and discharge disagreements. The chatbot responds to general queries about the occupational risk system and its various actors, as well as providing personalised advice based on individual user circumstances.

During development, the SRT employed a focus group methodology using internal employees to interact with the chatbot and systematically improve its performance. This iterative testing process enabled the team to double the chatbot's knowledge base from an initial set to 200 validated answers, generating 8,500 different question formulations. Most significantly, this process reduced the margin of error from 30 per cent to 7 per cent, achieving a 93 per cent accuracy rate in Julieta's responses. The knowledge base has continued to expand substantially since launch, growing from the initial 500 to 600 answerable questions to over 3,500, incorporating interconnections with external sites, videos, and integrated applications.

When first launched, Julieta handled slightly over 300 daily consultations. By 2019, the chatbot's daily volume had grown to over 1,000 inquiries, with more than 85,000 users generating 200,000 exchanges and 500,000 annual consultations over the year. At its peak, Julieta represented 35 per cent of total inquiries across all SRT communication channels, demonstrating significant channel shift from traditional phone-based customer service to the AI-enabled digital channel.

The chatbot received multiple awards and recognitions. In 2018, Microsoft recognised Julieta with an Innovation Award in the 'Conversational Agent' category for using artificial intelligence to address industry challenges and boost productivity. In December 2020, the ISSA awarded the SRT a certificate of merit with special mention for Julieta during the 'Good Practices — Americas 2020' competition, held as a virtual forum on 3 December 2020. The ISSA evaluation assessed 138 good practices from 30 institutions across 18 countries. At the same forum, the SRT also received recognition for its virtual intake desk (mesa de entradas virtual), a 24/7 platform accessible via 'e-Servicios-SRT' using CUIL/CUIT credentials for initiating labour proceedings. The SRT's Technical Manager Felipe Llorente presented the institution's AI methodology at the ISSA forum, alongside representatives from Brazil's National Social Insurance Institute and Uruguay's Social Security Bank.

Multiple SRT departments were involved in Julieta's development and ongoing operation, including systems management, administrative oversight, and customer service quality control divisions, reflecting an institutional approach to the chatbot's deployment rather than a purely technical initiative.

Classifications follow the DCI AI Hub Taxonomy. Hover over field labels for definitions.

Social Protection Functions

Implementation/delivery chain
Outreach/communications/sensitisation primaryCase management
SP Pillar (Primary) The social protection branch: social assistance, social insurance, or labour market programmes. Social insurance
Programme Name Sistema de Riesgos del Trabajo (Occupational Risk System)
Programme Type The type of social protection programme, classified under social assistance, social insurance, or labour market programmes. View in glossary Work injury and occupational insurance
System Level Where in the social protection system the AI is applied: policy level, programme design, or implementation/delivery chain. View in glossary Implementation/delivery chain
Automation Subtype For operational automation cases: (a) document processing and generative staff assistance, or (b) workload and resource forecasting. (a) Document processing and generative staff assistance
Programme Description Argentina's occupational risk insurance system regulated by the Superintendencia de Riesgos del Trabajo (SRT). The system provides work injury benefits, occupational disease coverage, and rehabilitation services through private Aseguradoras de Riesgos del Trabajo (ART, occupational risk insurers). The SRT regulates the system, oversees ART compliance, administers medical commissions, and handles worker inquiries and complaints.
Implementation Type How the AI output is produced: Classical ML, Deep learning, Foundation model, or Hybrid. Affects validation, compute requirements, and governance profile. View in glossary Foundation model
Lifecycle Stage Current stage in the AI lifecycle, from problem identification through to monitoring, maintenance and decommissioning. View in glossary Monitoring, Maintenance and Decommissioning
Model Provenance Origin of the AI model: developed in-house, adapted from open-source, commercial/proprietary, or accessed via third-party API. View in glossary Commercial/proprietary
Compute Environment Where the AI system runs: on-premise, government cloud, commercial cloud, or edge/device. View in glossary Commercial cloud
Compute Provider The specific cloud or infrastructure provider hosting the AI system. Microsoft Azure
Sovereignty Quadrant Classification of data and compute sovereignty: I (Sovereign), II (Federated/Hybrid), III (Cloud with safeguards), or IV (Shared Innovation Zone). View in glossary IV — Shared Innovation Zone
Data Residency Where the data used by the AI system is stored: domestic, regional, or international. View in glossary Not documented
Data Residency Detail Additional detail on the specific data hosting arrangements and jurisdictions. Built on Microsoft technology; specific data hosting location not documented in available sources
Cross-Border Transfer Whether data crosses national borders, and if so, whether documented safeguards are in place. View in glossary Not documented
Decision Criticality The rights impact of the decision the AI supports. High criticality requires HITL oversight; moderate requires HOTL; low may operate HOOTL. View in glossary Low
Human Oversight Type Level of human involvement: Human-in-the-Loop (active review), Human-on-the-Loop (monitoring), or Human-out-of-the-Loop (periodic audit). View in glossary HOTL
Development Process Whether the AI system was developed fully in-house, through a mix of in-house and third-party, or fully by an external provider. View in glossary Mix of in-house and third-party
Highest Risk Category The most significant structural risk source identified: data, model, operational, governance, or market/sovereignty risks. View in glossary Model-related risks
Risk Assessment Status Whether a formal risk assessment, informal assessment, or independent audit has been conducted for this system. Informal assessment
Documented Risk Events No publicly documented adverse events. Initial error margin of 30% was reduced to 7% through iterative testing with employee focus groups before achieving production-level accuracy.

Risk Dimensions

Operational and system integration risks

Impact Dimensions

Autonomy, human dignity and due process
  • Data minimisation controls
  • Human oversight protocol
CategorySensitivityCross-System LinkageAvailabilityKey Constraints
Beneficiary registries and MISPersonalSingle source (no linkage)Currently available and usedSRT worker database containing case file records (expedientes), ART assignments, medical commission assignments, and claim status information; accessed via identity verification to provide individualised responses
Unstructured and text-based contentNon-personalSingle source (no linkage)Currently available and usedKnowledge base of over 3,500 curated question-answer pairs covering occupational risk procedures, regulations, and system information; expanded from initial 500-600 through iterative focus group testing generating 8,500 question formulations

100% Seguro (2018) 'Inteligencia Artificial: premian a la SRT por \"Julieta\"', 100% Seguro, 2018. Available at: https://100seguro.com.ar/inteligencia-artificial-premian-a-la-srt-por-julieta/ (Accessed: 26 March 2026).

View source News article / media

Boletín Digital de ART (2018) 'ARGENTINA | Presentamos a \"Julieta\", el Chat Bot de la SRT', Boletín Digital de ART, 14 June. Available at: https://boletindigitaldeart.wordpress.com/2018/06/14/argentina-presentamos-a-julieta-el-chat-bot-de-la-srt/ (Accessed: 26 March 2026).

View source News article / media

ISSA (no date) 'Julieta Lanteri, primer chatbot de la administración publica nacional argentina', ISSA Good Practice gp/198017. Available at: https://www.issa.int/gp/198017 (Accessed: 26 March 2026).

View source Other

Argentina.gob.ar (2018) 'Presentamos a \"Julieta\", el Chat Bot de la SRT', Argentina.gob.ar Noticias. Available at: https://www.argentina.gob.ar/noticias/presentamos-julieta-el-chat-bot-de-la-srt (Accessed: 26 March 2026).

View source Government website / press release

Argentina.gob.ar (2020) 'La SRT premiada por la AISS en el concurso \"Buenas Prácticas — Américas 2020\"', Argentina.gob.ar Noticias, December 2020. Available at: https://www.argentina.gob.ar/noticias/la-srt-premiada-por-la-aiss-en-el-concurso-buenas-practicas-americas-2020 (Accessed: 26 March 2026).

View source Government website / press release

Argentina.gob.ar (2020) 'Foro virtual AISS: La SRT expuso sobre inteligencia artificial', Argentina.gob.ar Noticias, December 2020. Available at: https://www.argentina.gob.ar/noticias/foro-virtual-aiss-la-srt-expuso-sobre-inteligencia-artificial (Accessed: 26 March 2026).

View source Government website / press release
Deployment Status How far the system has progressed into real-world operational use, from concept/exploration through to scaled and institutionalised. View in glossary Full Production Deployment
Year Initiated The year the AI system was first initiated or development began. 2018
Scale / Coverage The scale and geographic or population coverage of the deployment. Nationwide; over 85,000 users generating 200,000 exchanges in 2019; handles 35% of total SRT inquiries across all channels; knowledge base expanded from 500-600 to over 3,500 answerable questions
Funding Source The source(s) of funding for the AI system development and deployment. SRT operational budget (Argentine government)
Technical Partners External technology vendors, academic partners, or development partners involved. Microsoft Argentina (NLP technology, development collaboration, hosting)
Outcomes / Results Error margin reduced from 30% to 7% (93% accuracy) through focus group testing methodology. Knowledge base expanded from 500-600 to over 3,500 answerable questions. Daily consultations grew from 300 at launch to over 1,000. Handles 35% of total SRT inquiries across all communication channels. In 2019, processed 200,000 exchanges from more than 85,000 users. Received Microsoft Innovation Award 2018 (Conversational Agent category) and ISSA Good Practice certificate of merit with special mention (Americas 2020).
Challenges Initial 30% error rate required significant iterative improvement through employee focus groups before achieving acceptable accuracy. Continuous expansion of the knowledge base is needed to address new query types and procedural changes. The chatbot handles informational inquiries and case status checks but does not make determinations on benefits or eligibility, limiting its scope within the broader occupational risk process. Data residency and cross-border transfer implications of using Microsoft cloud infrastructure are not publicly documented.

How to Cite

DCI AI Hub (2026). 'SRT 'Julieta' Chatbot for Occupational Risk Inquiries', AI Hub AI Tracker, case ARG-002. Digital Convergence Initiative. Available at: https://socialprotectionai.org/use-case/ARG-002 [Accessed: 1 April 2026].

Change History

Updated 1 Apr 2026, 08:11
by system (system)
Created 30 Mar 2026, 08:38
by v2-import (import)