Pfizer’s commercial activities currently generate enormous volumes of unstructured text from stakeholders in the healthcare eco-system. As data volumes continued to increase, human-centric mining of this unstructured data was becoming more difficult, extremely time consuming, and cost prohibitive. Pfizer needed a solution that would enhance its ability to quickly respond to feedback and proactively identify signals from the marketplace.
To address this challenge, Pfizer set out to apply cognitive computing techniques, namely text mining, natural language processing, and machine learning algorithms to revolutionize its ability to analyze and process the unstructured data collected across various channels.
The resulting visualizations allow the user to explore unstructured data from both a “top-down” and a “bottom-up” approach. The bottoms-up approach is particularly intriguing, as the company is finally able to let the data speak to discover hidden insights.
Pfizer’s cognitive analytics platform enables users to uncover patterns or trends in the unstructured data it collects. Currently, examining these trends allows business users to uncover insights that inform Pfizer’s medical communication strategies as well as day-to-day operational improvements, benefitting patients and shareholders.
Internally, Pfizer has completed a pilot and is beginning to build these same capabilities for delivery of a cognitive-enabled process for the monitoring of its promotional activities – specifically to ensure that internal and external communications meet the high standards of Pfizer’s policies as well as those of the industry’s regulatory bodies. Initial results demonstrate the potential to shorten the company’s review efforts by up to 85 percent versus current human-centric approaches.