Built a WhatsApp first AI appointment management product for healthcare

Challenges & Solutions

1. Identifying Customer Intent based on WhatsApp Chat:

One of the initial problems that arrises in building chat based systems is identifying the intent of the patient. The patient can say any sentence and in any language or format. Identifying the correct intent based on that message will accordingly allow the patient further actions. For MediSync, it was identifying intent of Booking, rescheduling, cancelling or general inquiries.

What We Did: We built an intent agent which was fine-tuned to identify the intent of the user. The intent to be identified was of booking, rescheduling, cancellation, or general query intent. Whenever a new user input is received, this agent identifies the user intent and accordingly redirect the user flow to respective flow.

2. Identifying the right doctors based on symptoms



For any healthcare company, they get patients in wide range of educational backgrounds, age, city and more. The system needs to work robustly for every type of patient. Many times, patient asks narrate symptoms and ask for an appointment. In a multi-speciality healthcare system, this becomes a challenge. We built a specialised routing agent which routes patients to specific doctors based on their symptoms. 





What We Did: We built a special agent which was fine-tuned with data of symptoms, co-occuring symptoms and respective specialisation mapping. Based on user’s symptoms, the agent would identify the specialisation and within which specific doctor would be aligned.

Results

Reduced operational overhead by 60%
Improved outreach and customer satisfaction
Return on Investment at 3.5x

Conclusion

MediSync is a powerful tool for the new generation of technological revolution. It reaches patients where they already and makes it easy for them to engage with the healthcare brand