How Rude! Computer Scientists are Building an Enraged Robot

A computer robot is analyzing calls from angry customers so companies can make changes in their customer service procedures. JosefKubes/Brian A Jackson/Thinkstock
A computer robot is analyzing calls from angry customers so companies can make changes in their customer service procedures. JosefKubes/Brian A Jackson/Thinkstock

Its brain is being filled with reams of angry tirades — hundreds of millions of them — culled from the calling center files of one of Australia's main banks. The venomous outbursts are customer interactions gone awry, and it will be Radiant the robot's job to analyze them and figure out how to use that collective anger to prevent more of the same in the future.

New Zealand's Touchpoint Group, a customer engagement consultancy and technology firm, is the force behind Radiant, investing more than NZ$500,000 (around U.S. $350,000) in this "big data" analytics software project. (Big data is an IT buzzword for massive volumes of data that are so large it's difficult to process them with traditional software techniques. But successfully analyzing big data can reveal patterns and trends, namely those relating to human behavior and interaction.)

"Using a combination of text analytics, natural language processing and machine learning, [Radiant] pinpoints and explains ... the top root causes of poor interactions with the contact center," says Mark Thompson, Touchpoint chief operating officer, via email.

Here's how it will work. As Radiant listens in on a company's call center calls, the software will automatically analyze all of the incoming data in innumerable combinations to examine how customers are interacting with various products, policies, procedures and even other people, before blowing their lids.

Then it will determine why some customers lost it, plus suss out any triggers behind their outbursts. The top three problematic issues — those that are most damaging to the company's brand and bottom line — will show on a dashboard, allowing the company to take action to combat them. The actions might be changes in a call center's operating procedures, for example, or incorporating specific bits of product knowledge in the training process.

"Essentially, it provides the 'ambulance at the top of the cliff' rather than at the bottom," says Thompson. 

But potential company changes won't be limited to call centers, as the root cause of customer anger may actually be an online banking experience, for example, or an ATM incident, app malfunction or even an experience at a retail branch. Figuring out the real reason why customers are angry is what makes Radiant so valuable. Thompson says there's yet another value to solving issues that vex customers — the more problems a company solves, the fewer complaints that will flow into its call center, and the faster employees can answer calls. And frustration with long wait times is often a major issue with call centers.

The first commercial product based on Radiant's ire will release in early April, according to Thompson, targeted at the banking and financial services market. It currently has no name or price tag. At least one more product release will occur later in 2016.

Down the road, the company hopes to expand Radiant to other industries which deal with customer service complaints. The Touchpoint Group will also offer a customization service for companies who will need Radiant to recognize words specific to its particular customer base.

Radiant's content analytics software is not a new concept. The same basic premise lies behind IBM's Watson, the question-answering computer system that beat Jeopardy contestants in 2011.