Ahead of next week’s To Bot or Not Webinar, Deon answered a few questions about the future of AI and Customer Support.
1) What, to you, is the difference between Automation and Augmentation in Customer Support?
Automation is typically about getting rid of repetitive tasks, and often has a transactional nature. Augmentation is about keeping the human (both the agent as well as the customer) front and center. It’s not about trying to replace people but about giving them the tools they need to be more efficient, find information faster, and ultimately have a more delightful day.
2) What is the biggest myth in AI and Customer Support that you would like to dispel?
That you need to manually label your data, or have “clean” knowledge base articles, before you can start using AI. We often see Support Leaders say, “we know we need to use AI here but are afraid to start because we know our knowledge articles [or our taxonomy] need to be updated.”
Thanks to modern advances in natural language understanding, and modern machine learning techniques like unsupervised learning and transfer learning, there are many novel ways to extract implicitly “labeled data” from existing sources. For example, every time an agent attaches a link to an article in a response to a ticket, that is an implicit label that this article can be used for this kind of question. And it usually happens naturally in the workflow.
3) What is the most memorable Customer Support interaction you’ve ever experienced?
When I called Brex because my card wasn’t working. Surprisingly they picked up in under a minute at like 10pm, the agent was super cheerful, respectful, and helped me diagnose a bug in their system. I got my new card issued then and there and was off to the races. Thanks Brex!
