FCAT: Customer service, conquered

You know the drill. You call, get put on hold, finally reach a human, only to get transferred. But at Fidelity, that’s just not good enough. So we at Labs took on the task of helping the Fidelity contact center elevate its customer experience.

Faster service is better service

We saw an opportunity to improve the contact center experience by solving customer issues faster, and thereby reducing their level of frustration. To do this, we had to get each caller to the right representative with the right skills while minimizing transfers.

Maybe data science can help

To get smarter about routing calls through the automated system, a handful of representatives were listening in on customer calls and filling out a survey after each one. Data scientists would then analyze the surveys to create an algorithm to predict why customers were calling. Fidelity would use the algorithm to improve the menu customers got when they called in. But the algorithm initially had low accuracy.

Instead we challenged the representatives to beat the algorithm. Drawing on their personal expertise with customers, could they predict caller issues any better? They stopped listening to the calls and just looked at the data. Turns out the accuracy of their predictions initially also was low.

The representatives learn to innovate

Because the representatives had the most insight into the call center experience, we decided the solution had to come from them. So we taught them some innovation techniques including Design Thinking, and Lean, put them in a room for six weeks, and waited to see what they would come up with.

On their own, the representatives honed their approach. They listened to more than 3,000 calls and mapped out the multitude of call types and predictors. They relayed the information back to our data scientists, who then updated the predictive model. Accuracy in predicting the reason for a call soared, which was great. But we still weren’t at 100%.

What if we simply asked people to tell us why they’re calling?

The representatives tested a new idea by answering calls and pretending they were the automated system. They asked the caller to state the reason for their call. Then they handed the phone to the appropriate representative who was sitting nearby. As they passed the phone, they whispered the reason for the call so the next representative knew what to expect.

And so the Whisper was born

With the Whisper, callers are asked by the automated system to describe why they are calling in their own words. The system then records their description, and their exact words are routed to the best-suited associate, who hears the Whisper before answering.

Our simple whisper echoed loudly

Adopting the Whisper meant fewer call transfers and a better overall call experience for customers. It also created better training opportunities for call center representatives—an added plus.

A typical phone tree experience.

Customers simply stating why they’re calling.



Patent Fest: Unlocking Creativity

Read more  


We average 12 patents a year.

We created the first investing app for a smartwatch.

We make prototypes with everything from pom poms to fly swatters.

Our work spaces are flexible and ever changing.

We have offices in the US, India, Ireland, China, and the UK.

We own 2 of the first 50 NVIDIA DGX-1 artificial intelligence supercomputers.