The Departmental Houses for Disabled People (MDPH) face major challenges: influx of requests, complexity of files, processing delays... To address these, the Marseille-based start-up Provence.ai (formerly Hyperbios) offers them an on-site AI platform that automates document recognition, data extraction, and dossier verification. The goal: to reduce processing times in line with government objectives while easing the workload of territorial agents.
A chronic bottleneck
By the end of 2023, more than 6 million people, nearly 10% of the French population, relied on the MDPH to obtain essential aid, having at least one open right. In the same year, nearly 1.8 million new applications for opening or re-examination were filed, an increase of 9% compared to 2022. Despite the teams' efforts, processing times reached an average of 4.8 months by autumn 2024, compared to the 2 months targeted by the government.
The solution developed by Provence.ai automates the analysis of supporting documents (medical certificates, forms, identity documents), compliance verification of files, and detects possible fraud. Its operation on departmental IT infrastructures ensures compliance with digital sovereignty and confidentiality requirements, major concerns in the processing of sensitive medical and personal data.
Tangible results
With a reported reliability of 99.7%, the AI, deployed in more than 15 departments, has reduced processing times from several months to a few weeks and increased team productivity by 30%.
For the founders of Provence.ai, the driving force of the project remains deeply human. Nicolas Woirhaye states:
"Provence.ai was born from our personal commitment and life stories marked by waiting and administrative complexity. Beyond performance, accelerating the processing of disability files is a human imperative for more than 6 million citizens and for the thousands of agents dedicated to public service."