Intelligent objection handling app for medical representatives

The medical and pharmaceutical industry is growing rapidly thanks to the development of scientific knowledge, access to new technologies, strong competition as well as changes in legal regulations. However, the medical environment is very complex and demanding, creating a need to deepen the expertise of medical representatives.

The sales of medical products to physicians and other healthcare professionals requires not only knowledge of trade, but also advanced medical expertise.

To facilitate the work of medical representatives selling oncological medications, for one of our clients, deepsense.ai created an application that uses advanced data analysis to propose the individual selection of sales and medical arguments tailored to a physicians’ predicted concerns and questions. Based on data collected from various sources, such as the location of the physicians’ practice, specialization, speeches at scientific conferences, scientific publications and other publicly available information, a profile of a given physician’s competences was created. Based on this profile, machine learning models predicted the set of questions and objections that this doctor would be likely to have. The solution operates on the notion that a given healthcare professional would express reservations similar to those previously raised by others from their social circle, including key opinion leaders. The measure of similarity between physicians was used within the k-nearest neighbour regression framework.

Apart from predicting the reservations, the application was equipped with natural language processing capabilities to extract and present the topics trending in social media and insights from recent oncology research.

The entire solution was deployed in AWS cloud, while the objection handling model was developed in AWS Sagemaker.

The solution improved the efficiency of medical representatives who, before meeting the physician, could analyze their profile and prepare to address potential questions.