Each colour maps a published value to the exact source passage used to support it. Only bounded excerpts are public; administrators can inspect the complete captured source.
Title
Doctolib builds an AI consultation assistant on Azure OpenAI Service and Mistral Large that summarises consultations in 15 seconds
…ny leverages Microsoft technologies to develop an AI-powered medical assistant, integrating both Azure OpenAI Service and Mistral Large on Azure. Medical consultations are now summarized in 15 seconds, allowing practitioners to spend twice as much time with patients. READ THE STO…
Description
Doctolib, a French eHealth company, built a consultation assistant using Azure OpenAI Service (GPT-4o) and Mistral Large on Azure, fine-tuned on proprietary French-language data. It transcribes consultations in real time and produces a structured summary within 15 seconds, which the practitioner reviews and validates. Audio and transcription are destroyed soon after validation. It was beta-tested with over 400 volunteer health professionals.
…e details Explore related stories Contact Sales Contact Sales Executive summary As one of the leading eHealth companies in France, Doctolib empowers healthcare practitioners to focus on what they do best: patient care.…
…osoft’s expertise and advanced AI technology to power its conversational agent. The assistant transcribes the consultation in real-time and turns it into a structured, relevant, comprehensible summary within only 15 seconds. Practitioners, on the other hand, will carry out the final step—carefully revie…
…mprehensive Azure AI Studio platform. To ensure the highest degree of accuracy, all these models were fine-tuned on proprietary data reannotated in-house and in the native French language. “There are a lot of controls from a data and AI point of view on the model prec…
…for their work, they are reconnecting with their vocation,” explains Talpaert. Both the audio and the transcription are destroyed soon after the medical summary is checked by the practitioner who remains in the driving seat. A secure, compliant, and tailor-made applications to handle sensitive medical d…
…trusted: a consultation assistant endorsed by hundreds of health professionals The company already beta-tested the assistant with over 400 voluntary health professionals over thousands of consultations per week, and received extremely positive feedback. The feedback from health professiona…
Company
Doctolib
Doctolib transforms consultations and doctor-patient relationships with Azure AI | Microsoft Customer Stories This is the Trace Id: b19750804d05d424c96d8d05dcaf…
Country
France
…ure OpenAI Service SERVICES AND SUPPORT ORGANIZATION SIZE 1,000-9,999 employees COUNTRY France BUSINESS NEED Customer experience Digital service Employee experience Artificia…
Industry
Healthcare
…e details Explore related stories Contact Sales Contact Sales Executive summary As one of the leading eHealth companies in France, Doctolib empowers healthcare practitioners to focus on what they do best: patient care. The company leverages Microsoft technologies to develop an AI-powered medical a…
Problem
Taking notes on the computer for the patient file and putting together the medical record is time-consuming for practitioners, who Doctolib wants to be able to spend more time looking at their patients rather than their computer screen.
…computer screen,” explains Arthur Talpaert, Lead Product Manager at Doctolib. “What’s time-consuming for the practitioners is taking notes on the computer for the patient file and putting together the medical record.” With this ambition at heart, Doctolib leveraged Microsoft’s expertise and adva…
Solution
Doctolib built a Consultation Assistant that transcribes the consultation in real time and turns it into a structured summary within 15 seconds. Practitioners review, edit and validate the information before it is integrated into the patient's file or shared with the patient. The assistant is based on a series of models including GPT-4o on Azure OpenAI Service, chained with Mistral Large on Azure and built on Azure AI Studio; the models were fine-tuned on proprietary data reannotated in-house in native French. Both the audio and the transcription are destroyed soon after the summary is checked by the practitioner.
…g the different models on Azure—and the level of data training personalization. The company chained different models—from Azure OpenAI Service to Mistral Large on Azure —and built the AI app responsibly through the comprehensive Azure AI Studio platform. To ensure the highest degree of accuracy, all these models were fine-tuned on p…
…it into a structured, relevant, comprehensible summary within only 15 seconds. Practitioners, on the other hand, will carry out the final step—carefully reviewing, editing, and validating medical information to be integrated directly in the patient’s file or shared with the patient. Based on a series of models including GTP-4o, available on Azure OpenAI Service…
Business value
According to the Microsoft customer story, consultations are summarised in 15 seconds, aiming to let practitioners spend twice as much time interacting with their patients, and physicians report a reduced mental workload. The assistant was beta-tested with over 400 voluntary health professionals over thousands of consultations per week, with extremely positive feedback.
…ing practitioners to spend twice as much time interacting with their patients. “Physicians share their views that this is a major contributor to reducing their mental workload. They’ve regained a sense of appreciation for their work, they are reconnecting…
…s. Improving patient care with Azure OpenAI Service and Mistral Large on Azure “Our mission is to help health professionals spend twice more time looking their patients in the eye rather than their computer screen,” explains Arthur Talpaert, Lead Product Manager at Doctolib. “What’s time-consu…
…trusted: a consultation assistant endorsed by hundreds of health professionals The company already beta-tested the assistant with over 400 voluntary health professionals over thousands of consultations per week, and received extremely positive feedback. The feedback from health professionals and patients is incredible, both in term…
AI capabilities
Generative AI, Large Language Models
…ny leverages Microsoft technologies to develop an AI-powered medical assistant, integrating both Azure OpenAI Service and Mistral Large on Azure. Medical consultations are now summarized in 15 seconds, allowing practitioners…
Technology
Azure OpenAI, Mistral Large
…ny leverages Microsoft technologies to develop an AI-powered medical assistant, integrating both Azure OpenAI Service and Mistral Large on Azure. Medical consultations are now summarized in 15 seconds, allowing practitioners…
Deployment model
Cloud
…aspects from a technical point of view,” Talpaert says. The company also needed flexibility when it came to building and deploying the different models on Azure—and the level of data training personalization. The company chained different m…
Deployment options
cloud
…aspects from a technical point of view,” Talpaert says. The company also needed flexibility when it came to building and deploying the different models on Azure—and the level of data training personalization. The company chained different m…
Headline outcome
…nversational agent. The assistant transcribes the consultation in real-time and turns it into a structured, relevant, comprehensible summary within only 15 seconds. Practitioners, on the other hand, will carry out the final step—carefully revie…
