We are living in a time where technology is advancing rapidly, and many industries have the opportunity to leverage this transformation to improve what they do. One of the sectors with the greatest potential for a significant leap forward is, without a doubt, healthcare. Among the technologies generating the most expectations are AI agents, systems based on artificial intelligence capable of supporting and enhancing physicians’ work. 

An AI agent is a system based on artificial intelligence that, unlike ChatGPT which responds when prompted, can act autonomously to achieve a goal. It does not wait for instructions. It perceives its environment, processes information, and decides what actions to take to move toward its objective. The ability to define different roles and how they interact with each other makes it an extremely powerful tool. 

Today, a physician typically has 15 to 20 minutes per consultation to address a clinical situation. Within that limited time, they must review the patient’s medical history, listen to the patient, explore possible diagnoses, determine appropriate treatments, and finally document the consultation in the system. 

What if, instead of facing this challenge alone, physicians could rely on a team of virtual colleagues powered by AI? In other words, multiple agents based on large language models (LLMs), each with a defined role and specialization, contributing knowledge based on their training. Some potential roles could include: 

• Medical history analyst: quickly processes the patient’s history, highlighting relevant information that might otherwise be overlooked during a standard consultation. 
• Diagnostic support: suggests possible diagnoses, complementary tests, and helps rule out conditions based on up-to-date clinical data. 
• Interviewer: proposes key questions to guide the consultation, helping ensure that no important topics are left unexplored. 
• Specialists: represent expertise across different medical fields such as cardiology, diabetology, or neurology, providing tailored recommendations or alerts based on the patient’s profile and available data. 

In this model, the physician remains at the center, applying their expertise, judgment, and knowledge, but now supported by a team of virtual colleagues. 

Unlike earlier stages, such as November 2022, current models allow for much greater flexibility when discussing medical topics. Initially, with ChatGPT, it was necessary to carefully craft prompts to navigate content limitations. For now, that constraint appears to have diminished. 

Thanks to these advances, language models can now become a valuable tool for healthcare professionals by delivering faster, clearer, and more instruction-aligned responses. 

For these possibilities to translate into real benefits, several challenges must be addressed. It is essential to protect patient privacy, ensure transparency in how these agents operate, and guarantee that technology always supports the physician’s judgment rather than replacing it. 

In 2025, we continue to see strong competition between open-source and proprietary AI models, with different sizes, speeds, and approaches. It is reasonable to expect that we will soon have models specifically trained for healthcare, adapted to local contexts. 

There is also an often overlooked advantage. Physicians are already trained to anonymize clinical data, which allows them to consult with human or virtual colleagues without compromising patient confidentiality. This opens the door to more fluid collaboration with these intelligent assistants. 

The successful implementation of AI agents in healthcare requires strong data governance, with clear policies on data collection, protection, and usage. This ensures privacy, quality, transparency, fairness, and security, as well as informed patient consent, medical accountability in automated decisions, and ethical, bias-free adoption. 

The integration of AI agents in healthcare creates the opportunity to transform each consultation into a true collaborative effort. Not to replace physicians, but to empower them with tools that provide real-time data, insights, and recommendations. 

Imagine a consultation where every professional is supported by a virtual team that enhances their decision-making capabilities. The challenge and the opportunity are both on the table. 

At Quanam, we believe technology can improve people’s health and make the work of healthcare professionals more effective. This journey is just beginning, and we aim to be an active part of that transformation. 

Kenji Nakasone

Lead AI Innovation Engineer