While Doctolib has established itself as the most visible online interface, the reality is that it only covers 30 % of practitioners and rarely all the reasons for consultation. 70 % of meetings still go through the phone, a saturated channel where secretaries have to face patient reception, many administrative and incessant calls.
Vocca.ai, founded by Elliot Offenberg and Hugo Danet, is positioned on this weak link with a vocal AI specialized in the medical sector. Available 24/7, it supports all calls and integrates directly into business software. Unlike a generic solution, AI is trained in medical terminology and workflows specific to each specialty. It can manage complex cases, such as the combination of exams (MRI and radiography on the same niche), while taking into account contraindications (pacemaker, claustrophobia).
Natural adoption on the patient side
The first returns show that AI is perceived as a natural extension of the secretariat. During a test, 40 % of patients did not identify in the first minute that they dialogue with a machine. The rate of call resolution, which capped at 20 % in the creation of the company, today reaches 80 %. Partner establishments note a coverage of 100 % of calls, without waiting time, and a significant drop in “no-shows” thanks to automated reminders. Patients highlight permanent availability and the possibility of taking the time during the call, where a human secretary is often constrained by the workload.
Redefine the role of medical secretaries
Far from eliminating the secretariat, the solution redistributes the missions. The AI absorbs standardized and administrative requests (invoices, referrals, confirmations), while the secretaries focus on physical reception and complex situations. The tool offers a supervision interface which notifies the secretaries in the event of an emergency or a particular request, leaving them the last word on critical decisions. For practitioners, the equation is attractive: the outsourced remote school is one of the three main spending stations, after rent and payroll.
Multi-agent technological architecture
Technically, Vocca is based on the most advanced language models (LLM), structured in the form of specialized agents. Each agent is drawn to a case of specific use: making an appointment in ophthalmology, management of patient files, transcription of medical vocabulary. The team of engineers worked on the reliability of transcription and the extraction of sensitive information, such as names and surnames, with a success rate of 99 %. The company has also developed safeguards to limit the “hallucinations” of the models, manually encoding hundreds of use cases in order to ensure operational consistency.
A tension market and a target of 10,000 practitioners
The company is targeting a vast market in France, 200,000 health professionals must manage a call flow daily that they cannot absorb. Already, 40 % of them resort to TELE -SECRETARIES, the cost of which can represent several tens of thousands of euros per year. VOCA aims to equip 10,000 practitioners by the end of 2026. The demand is there, but the company must strengthen its workforce to deploy on a large scale and improve its integrations with medical software. A high -growth sales team has been set up, going from 1 to 6 people in a few months, with a short -term doubling plan.
A quick fundraising
Vocca raised 4.7 million euros to support this acceleration. The rounding was led by Speedinvest (Austria) and Firstminut Capital (United Kingdom), both familiar with health and IA subjects, joined by several French business angels: Kima Ventures (Xavier Niel funds), the founder of Datadog, in particular. The lifting ended in record time, with first Term Sheets received in 48 hours. The funds will strengthen technical and commercial teams, extend software integrations and prepare the internationalization of the solution.
