Lupin - a multinational, pharmaceutical company wanted to create a high context, NLP (Natural Language Processing) enabled chatbot to educate and interact with patients around diseases like diabetes, hypertension and tuberculosis.
The chatbot had to -
- Work on a public, social messenger like Facebook Messenger to allow millions of users world-wide to interact with it.
- Answer more than 2000+ medically verified questions.
- Maintain context of a disease when a patient asked questions in her natural style and language. For example, when a patient asked a series of questions around diabetes, Anya had to recognise that the context was 'diabetes' and continue to maintain the context until the patient changed the subject of the conversation to another disease.
- Offer logical suggestions to patients about the next probable questions, using AI (Artificial Intelligence) and ML (Machine Learning).
- Work in multiple languages like Hindi and Telugu in addition to English.
- Throw insights and analytics for Lupin team to understand what the patients were curious or anxious about for a disease.
"Chatbot as a technology has enormous potential to aid in medical assistance and to support in building awareness and breaking myths associated with various therapies." - Rajeev Sibal, Lupin IRF president
Quickwork created Anya chatbot which met above requirements, by integrating and connecting multiple applications as follows -
- Facebook Messenger as messenger to interact with patients
- Dialogflow to train Anya in NLP (Natural Langugage Processing) queries
- Quickwork Context Manager to maintain high context in conversations
- Amazon Machine Learning engine for automated, self-learning of Anya
- PredictHQ to predict next set of probable questions
Anya became very successful with more than 9,00,000 users asking questions over multiple sessions, 3 languages with 92% efficacy in terms of correct and contextual content delivery.
As seen on news:
Build your high context NLP chatbot on Quickwork Platform, quickly.