Artificial Intelligence in Your Bio-Pharma Contact Center and Where to Use It

As the bio-pharma contact center market learns more from experimenting with Artificial Intelligence (AI), Centerfirst wanted to weigh in with our insights on exactly where AI tools currently are best used to improve the efficiency, accuracy, and customer experience delivered in the contact center.

Artificial Intelligence

Let me start by saying the AI tools aren’t as good as the sales demo…at least not without a lot of post-purchase support and customization (think Siebel in early 2000). That doesn’t mean they shouldn’t be explored, it just means that their value as a disintermediating force in bio-pharma contact centers is still a few years away.

Here’s why. There are three hurdles that AI needs to overcome in the bio-pharma contact center to have a wide-ranging impact on analytics, insights, and customer experience.

  1. Voice to text technology is improving but still only between 80% and 90% accurate even by the most robust of automated translators. It’s important to also understand how these percentages are calculated. Don’t think AI translates 90% correct, think AI can be up to 90% accurate on certain data sets.
  2. AI’s ability to comprehend translated text reduces accuracy even further. Comprehension is difficult in any language, even for human natural language speakers. Comprehending an imperfectly translated text even by the knowledge of robots leads to inaccurate outcomes.
  3. Because terms used in bio-pharma and healthcare are more uncommon and complex, and the standards of accuracy are higher, bio-pharma and healthcare should do what they do best and experiment with certain parts of AI before committing large resources.

How should you use AI today?

Let’s segment “usage” into the three categories.

  1. Use Frequently – Text miner – These tools are terrific and ideal for recognizing words or phrases that are used repeatedly in text. We recommend turning these into word clouds and geo maps to help spot trends in activity, interest, and opportunity. Centerfirst uses text mining to QA our own work when monitoring for adverse events. One tip: The word “patient” was the most reliable for identifying potential adverse events.
  2. Use Sparingly – Voice to Text Translators – This technology has high promises but currently low ROI. The one area of bio-pharma that it may have some current application is in inside sales/sales support. Complications with translation compounded by the difficulties of comprehension in healthcare and the added expense of customization make the cost of the output much too high for the reliability. We continue to experiment with voice to text and will be the first to fully adopt it, but until then, we would recommend using it sparingly.
  3. Avoid – Reliance on translators for regulatory compliance – There will be a time when translators can be used for the assurance of regulatory compliance (i.e., identification of adverse events) but that time has not yet come.

Centerfirst has monitored hundreds of thousands of healthcare interactions using available current technologies and highly trained quality monitoring specialists. Let us help you develop a monitoring program that delivers the efficiency, accuracy, and customer experience you demand in your contact center.