We are starting work on a new research project called PulseDocs AI, that introduces an innovative model combining advanced scheduling automation, speech recognition, and artificial intelligence technologies to optimize the doctor-patient dialog. The proposed model is structured into three main components, each with a distinct role in improving the operational workflow in clinics and medical practices.
Technologies
- Programming Language: TypeScript
- Technologies/Frameworks: Medical LLM trained on Open AI Whisper and GPT-4o
- Database: Internal ERP
- Components: IVR, Open AI Whisper, GPT-4o
Objectives
- Reduce clinic costs by eliminating call centers and streamlining physician time
- Increase patient satisfaction by providing a simple, efficient and fast system
- Integration with the clinic’s database for automatic medical record saving and report generation
- Integration of an IVR system for automatic call handling
- Voice to text translation for better medical personnel time management
- Text refinement, automatic extraction of medical structure and generation of clinical documentation (e.g. medical letters, prescriptions)
How Will It Work
The system will manage appointments automatically: patients call a landline, the IVR (Interactive Voice Response) collects basic information, and the app finds and confirms an available slot without staff involvement.
The app will send automated reminders on the day of the consultation, prompting patients to confirm or cancel. This helps clinics avoid unused time slots and better allocate resources.
During consultations, doctors will be able to dictate notes, which will be then converted to text through voice recognition and saved directly in the database. This should remove manual data entry and speed up documentation.
Overall, the application is meant to streamline communication, reduce administrative workload, and improve workflow inside clinics.
The Innovative Conceptual Model of PulseDocsAI
The PulseDocsAI application introduces an innovative conceptual model structured around three main components, each with a distinct role in improving operational flow in clinics and medical offices:
- ASR Module – Automatic Speech Recognition
The introduction of an AI-based voice-to-text translator will allow physicians to dictate their consultation notes, eliminating the need to manually enter data into the system. The model is self-training: the more it is used, the better it will learn the particular vocal inflections and speech patterns of the doctor.
- The NLP (Natural Language Processing) module for medical documentation via GPT-4o
This module will use an LLM (Large Language Model) for text refinement, automatic extraction of medical structure, and generation of clinical documentation (e.g., medical letters).
- Automatic appointment management module (IVR – Interactive Voice Response)
This module will streamline initial interaction with the clinic via an automated patient management system.
Conclusion
The PulseDocsAI conceptual model will redefine doctor-patient interaction through:
✅ Automating appointments, reducing waiting times.
✅ Eliminating repetitive work for assistants with automated reminder calls.
✅ Optimizing documentation with a voice translator, reducing administrative tasks.
With this approach, the app has the potential to revolutionize clinical efficiency and provide a high level of accessibility for patients, significantly improving healthcare experience.
Project Financing
This project is co-financed through the North-East Regional Program 2021-2027, Priority 1 North-East – A more competitive, more innovative region; Specific objective 1.1: Developing and increasing research and innovation capacities and adopting advanced technologies; Operation 3: Support for the development of SME innovation capacity; Call for projects: CALL NO. PR/NE/2025/P1/RSO1.1/1/3.3 – Proof-of-concept