Virtual agents not yet ready for primetime customer service


Call center chatbots help, but don't replace, humans

We hear it all the time from optimistic, brand-driven companies or the tech vendors that serve them: "We'll close our call center and save millions by replacing people with chatbots!"


Chatbots might be getting smarter by the day, and artificial intelligence might be light years ahead of where it was a decade ago, but relative to humans, it's barely "intelligent." It's going to be years before virtual agents are equipped with enough smarts -- combining natural language processing, AI and algorithmic responses -- to solve more than rudimentary questions and tasks such as password change requests, balance inquiries and order shipping status.

Wise customer service and support executives aren't looking to replace humans at their call centers with virtual agents, but rather to augment their workflows. If chatbots can resolve those rudimentary, common, oft-repeated customer inquiries as well as recognize more complex questions and route them to experienced human agents, that's the sweet spot for this technology. The result: Creating efficiency, reducing call volume, fielding customer inquiries from a variety of channels -- email, social media, text, etc. -- and monitoring those channels so that humans need to do less of it.

Chatbots can be especially effective for administering customer service for B2C retailers doing a high volume of simple transactions or for self-service for training, manuals or assembling straightforward products. But the key to using AI-powered chatbots effectively comes down to thoughtful customization of your call center workflow and proper training of the humans who will be "working with them."

The articles in this handbook will help guide your organization's chatbot implementation and offer some expert tips along the way.