Case study — Healthcare
Emerald Medical Center
Emerald is a member of the SOFMEDICA Group of Companies, which operates in the medical field and is oriented towards education. The company has been present in Europe since 1994, with offices in Romania, Greece, Hungary, Bulgaria, and Cyprus. With over 15 years of experience, Emerald has a team of over 200 employees.

The Challenge
In the fields of radiology, imaging, and ultrasound, doctors and nurses spent a significant amount of time manually inputting data into computers. SOFMEDICA Group tried multiple competitors in the speech-to-text industry but was not satisfied with the accuracy level they achieved in the medical field.
The company aimed to develop and implement a custom software solution to automatically convert medical dictations and clinicians' notes from audio recordings into text. The goal was to transfer this information into SOFMEDICA's IT systems, which automatically collect patient data and build a patient's medical record. This would include information such as analyses, scan reports, and medical consultations.
Emerald aimed to increase its competitiveness and differentiate itself in the market by implementing a highly accurate speech-to-text solution, tailored for the medical field. This would result in:
- An improved customer and clinician's experience.
- Increased operational efficiency, including faster execution.
- Cost reduction in the medium and long term.
The Solution
Vatis Tech required a minimum of 30 hours of data recordings featuring medical tests interpretation, scans, and clinical notes from various medical specialties. That was necessary to label and validate sufficient learning data, which helped calibrate the Romanian artificial intelligence algorithms used within the Vatis Tech platform.
Emerald, along with a team of clinicians involved in the project, went the extra mile for a month to record audio of simulated interpretations of medical lab tests. The effort aimed to cover all medical terminology and abbreviations.
Vatis Tech's AI algorithms were fine-tuned using radiology audio data over a period of 2-3 months.

We are very happy to move from writing the results manually to implementing an artificial intelligence-based program to streamline the process of issuing radiology scan results. As next steps, we hope to implement this software in the physician practice as well, in order to help medical teams with the clinical documentation, especially when patient time is essential for a correct diagnosis and for the most efficient therapeutic management.
The Results
The transcription of medical terminology improved significantly
Vatis Tech's AI algorithms have been calibrated on medical terminology used in radiology and imaging through sustained efforts over a period of 2-3 months, using audio data.
Medical transcription with over 95% accuracy rate
Emerald achieved higher accuracy in transcription, sometimes reaching close to 100%.
A higher degree of efficiency was achieved
Speech-to-text applications have alleviated clinicians from time-consuming documentation tasks, enabling them to focus on patient care. The speech-to-text solution provided by Vatis Tech was fast, accurate, and automatically transcribed audio files 10x faster than manual transcription. This saves time and makes workflows faster and easier.
Next Steps
Implement the speech-to-text technology developed by Vatis Tech in all of their newly opened clinics.

Vatis Tech tackles the challenge of speech-to-text in medical and radiological language by employing artificial intelligence. With a focus on accuracy, Vatis Tech surpasses other applications for the Romanian language, offering reliable transcription results. Radiologists can enhance their daily workflow by dictating reports that are immediately transcribed into medically precise, semantically coherent, and stylistically accurate texts. The efficiency of Vatis Tech promises to reduce the time required for writing radio-imaging examination reports, providing a welcomed solution for healthcare professionals.
More case studies
Teams running Vatis Tech in production

Up to 4x cheaper
than the solutions mediaTRUST had tried, with enhanced recognition on specific keywords.
Read case study
3 hours of manual work down to minutes
letting JURIDICE.ro handle large volumes of hearings and depositions with ease.
Read case study
~5x faster than a human
Hours of transcription time are reduced to minutes for Mercury Research.
Read case study
For engineers who read the docs before the marketing page
Read the documentation, try it free, tell us how it goes.
More from Vatis
Transcription Software
Learn more about how Vatis Tech helps teams and individuals streamline their transcription workflow.
Learn moreSpeech-to-Text API
Supercharge your apps with Vatis Tech's accurate, accessible and affordable Speech-to-Text API.
Learn moreAudio Intelligence
Extract actionable audio insights in minutes from your transcription API.
Learn more