LLMs and empathy
### A healthcare use case: booking diagnostic examinations
The introduction of Large Language Models (LLMs) into healthcare conversational workflows is fundamentally transforming the way patients interact with clinical services. In particular, the integration of empathetic conversational capabilities opens new opportunities to improve the patient experience, even in seemingly operational scenarios such as booking diagnostic examinations.
### Conversational empathy in LLMs
For an LLM, empathy is not an emotion but a simulated behavior: it is the model’s ability to recognize the emotional context implicit in a user’s message and respond with language that conveys attention, courtesy, and understanding. This is especially important in healthcare, where patients may feel anxious, uncertain, or under stress—even when they are simply trying to schedule an ultrasound or an MRI scan.
### The Halia case by Unicorn Apulia
A practical example of empathetic LLMs applied to healthcare is [Halia](http://www.halia.it/), the virtual assistant developed by Unicorn Apulia to support the booking of diagnostic examinations at healthcare facilities and imaging centers. Halia combines a conversational LLM with practice management systems (such as Claris FileMaker or regional CUP booking platforms) to guide users through selecting the appropriate examination, understanding any required preparation, checking appointment availability, and confirming their booking.
Halia’s distinguishing feature is its ability to maintain a reassuring, clear, and personalized conversational style, even when users provide incomplete information, make typing errors, or express uncertainty. For example, if a patient writes:
> “I’d like to book an MRI, but I’m not sure whether I need a referral.”
Halia can respond:
> “No problem. Some MRI examinations require a physician’s referral, while others do not. May I ask whether this MRI was prescribed by your general practitioner?”
rather than replying with a generic or rigid message.
This empathetic approach helps to:
– reduce appointment booking abandonment rates;
– increase patient trust in the service;
– minimize unnecessary escalations to human operators.
Although empathy in LLMs is simulated rather than emotional, it can deliver tangible benefits in digital healthcare processes. In use cases such as diagnostic examination booking—where operational efficiency must be balanced with the need to reassure and guide patients—an empathetic conversational assistant like Halia represents an effective, real-world example of how conversational AI can enhance healthcare services.
