Pharma customer service reps can’t just Google the answers to healthcare provider (HCP) questions about their company’s drugs. But now, at Biogen, they can. The pharma company, along with Lexalytics, created a custom internal search engine that customer reps use to quickly find answers about the company's portfolio of products.
The artificial intelligence system replaces assembled-over-the-years FAQs, text-heavy docs and educational materials reps might otherwise have to comb through for answers. It uses natural language processing so that the human operator can literally type in the exact words a physician is saying and get an answer back quickly.
RELATED: Amazon Alexa's first HIPAA-sanctioned skills open new doors for pharma
The benefits are twofold: better and faster customer service to satisfy HCPs and a much-needed support tool in the surging information demand on customer service reps. In the past, if an operator couldn’t find an answer within one minute, the call was escalated to a medical director—costly for Biogen and frustrating for HCPs.
“One person cannot know the answer to every single question to answer inbound inquires,” Keith Ho, director of customer focus and medical digital at Biogen, said. “Over time, medical information systems created a low-tech menu way of collecting Q&A or FAQ for inbound calls. That is not ideal, especially because the operator has to search word documents for answers. We realized that had to change.”
Biogen nixed going straight to machines in part because physicians prefer a human helper, but also because Biogen wanted to implement in steps to ease into AI customer service and learn along the way.
RELATED: If pharma looks slow to adopt AI, it's got good reason, expert says
The AI search platform may potentially ease Biogen’s customer service talent issues as well. In testing the system, Biogen did an experiment using both experienced and inexperienced reps. The result? Ho said the inexperienced operators who had never answered calls before had the same result as the experienced operator used to handling the HCP calls. That makes it easier to start new operators as well as manage temps for vacations and sick days.
The next steps will be to build out the system, Ho said. Biogen will explore more functions on the platform such as capturing metadata about the callers—recognizing them and their histories, for instance—and also investigate other areas at Biogen where natural language processing machine learning could be used.