Outsourcing Emotion With A.I. Customer Service

Danielle Beurteaux

For many consumers, customer service has at best an intimate relationship with frustration, if not outright anger. A Consumer Reports survey found that 57 percent of respondents get so frustrated with calling customer service that they hang up without getting their problems solved or questions answered. Some of the customer service situations consumers hate most: being disconnected, transferred too often, customer service reps who can’t help with a problem—the list of issues is long (and much too relatable).

This is in part because call center technology has historically focused on benefiting the company that uses it over its customers. So some companies are looking to Artificial Intelligence to improve the customer service experience.

Google recently announced two new Google Cloud tools that offer A.I. capabilities to retailers. Based on language analysis and speech-to-text conversions, the tools recognize angry language and can prioritize responses to those unhappy customers (which begs the question: Will customers figure this out and game the system?).

On the heels of Google’s announcement, retailer Macy’s signed on with IBM to help its shoppers using Watson A.I. technology. The pilot program, “Macy’s On Call,” is the retailer’s response to the fact that many shoppers use their phones while they’re in the store. Initially slated for 10 stores, the service will enable customers to text questions, which Watson will answer. Because it’s a learning program, Watson’s answers will adapt over time for each store.

Germany’s Paralmind is a startup that recently launched in the customer service A.I. space, and earlier this year, startup DigitalGenius raised $4.1 million in seed money. The company’s Human+AI platform, which analyzes call logs and chat transcripts, is designed for integration into existing CRM programs. It’s not meant to completely replace human reps—the first layer is the “intelligent” answer, but the second is to recommend the best response to reps. Its aim is to handle basic, repetitive questions to free up reps for more challenging issues.

This hybrid model might be the way forward, and not just because A.I. is still in its infancy. Answering even simple questions require multiple steps, and the intelligence learning phase takes time—at this point, more time than getting a human to answer.

Companies are already taking the hybrid route. One example is Atom Bank, which has incorporated A.I. into its mobile app. The plan is that it will take care of the straightforward questions, but anything more challenging will be transferred to its customer service center. Stationer Staples has a similar setup for its business customers.

The initial step might be A.I. that helps human customer service reps do their job better so that customers are happier with the entire customer service experience. Customers prefer interacting with humans, whether in the store or on the phone. A.I. technology can help improve the human-to-human relationship, but it isn’t going to replace it entirely anytime soon.

For more insight on how technology is changing the customer experience, see The New Face Of Customer Service.


Danielle Beurteaux

About Danielle Beurteaux

Danielle Beurteaux is a New York–based writer who covers business, technology, and philanthropy. Her work has appeared in The New York Times and on Popular Mechanics, CNN, and Institutional Investor's Alpha, among other outlets.