At Parallel, we focus all our efforts on one goal: helping medical coding teams code more accurately, more quickly, and with greater confidence. The Assistant is one of the tools we provide to help them improve their coding.

Why we developed the Parallel Assistant

PMSI coding is complex. For every diagnosis, medical coding and Medical Information Department teams must interpret the rules, assess the available information, and make precise decisions.

Yet the tools available to them do not fully meet this need. Some legacy tools, such as aideaucodage.fr, are uneasy to use. Others, such as ChatGPT or Claude, are powerful but do not meet the sector’s requirements for health data protection, control over data processing, and verifiable answers.

At Parallel, we already work extensively with large language models in a secure environment. It therefore seemed obvious that we should make this technology directly available to medical information teams through a tool designed around their needs and constraints.

This is why we created the Parallel Assistant: to help teams find information, make decisions with confidence, and better understand their activity—without ever replacing their expertise.

The Assistant’s main use cases

Three use cases are brought together in a single interface.

1. Get answers to your PMSI questions

The Assistant understands the medical context of your questions, draws on official ATIH reference documentation, and precisely cites the sources used.

2. Ask questions about the patient record

The Assistant searches the patient record for the information needed to answer your coding questions.

3. Analyse the coding performed by Parallel

The Assistant analyses the coding performed by Parallel to help you understand where your stays are being revalued the most, with results presented as tables and charts.

Three use cases, one interface: answer your PMSI questions, identify the right code, and analyse Parallel’s impact on your stays.

The Assistant’s key features

Answers directly linked to official documentation

The Assistant draws on ATIH references - including the methodological guide, coding guidelines, CoCoA, and billing rules - and cites the documents and excerpts used. This makes it possible to verify every answer before making a decision.

Security designed for health data

Data is pseudonymised before analysis. The models are hosted on HDS-certified servers, with no data retention. You can also share documents with the Assistant: they are pseudonymised before any analysis takes place.

Results that are easy to visualise

For activity analyses, the Assistant presents results as ready-to-use tables and charts. A simple question can be used to explore thousands of stays, identify trends, and quickly share findings with your teams.

The Assistant accelerates research and analysis; medical coding and Medical Information Department teams remain in control of every decision.

Try the Parallel Assistant

The Assistant is already available in your Parallel application and included in your existing plan.

Try it here: Vers l'Assistant