Dataprev's Isaac AI System

Source: Dataprev, 7 November 2019
URL: https://portal.dataprev.gov.br/5a-semana-de-inovacao-dataprev-apresenta-isaac-solucao-de-ia

Isaac is an artificial intelligence solution developed by Dataprev to streamline benefit analysis at Brazil's National Social Security Institute (INSS). Named after science fiction author Isaac Asimov, the system employs predictive algorithms for decision-making in social security benefit recognition.

Key Capabilities

Automated Processing: The system addresses a critical backlog, as "about 40% of benefits requested in July 2019 were denied" with an average analysis timeframe of 89 days.

Parallel Analysis: Isaac can simultaneously process thousands of cases, using machine learning to cross-reference multiple databases and deliver reliable results in assisted or fully automated modes.

Risk Assessment: The technology features adjustable risk criteria and analyzes historical data to forecast outcomes.

Supporting Technology

Prova de Vida (Proof of Life): A complementary biometric solution eliminating mandatory in-person verification. It cross-references facial and fingerprint data from Brazil's Electoral Court and National Transit Department, validating citizen status via smartphone photography.

Context


The presentation addressed demographic pressure: Brazil's population over 65 will grow from 9.2% currently to 25% by 2060, while approximately 16,000 INSS employees could retire by 2020 -- those currently responsible for manual case analysis.