Cogito Dials Up $5.5M for Software to Coach Customer Service Reps
As consumers, we’ve all had a frustrating experience (or several) talking with a customer service representative on the phone. Maybe the voice on the other end of the line was speaking too quickly, not really answering your question, or simply being rude.
For many businesses, the volume of calls coming in to their call centers—millions per day, in some cases—makes it impossible for managers to closely monitor each of those calls and guide customer service workers toward a better conversation that keeps the customers satisfied.
That’s something a Boston-based company, Cogito, is trying to solve with its voice-analysis software that provides a real-time assessment of how well customer service calls are going and suggests ways the call center worker can improve the dialog with the customer.
The MIT Media Lab spinout, co-founded by renowned data scientist Alex “Sandy” Pentland, has tested its behavioral analytics software in multiple settings, including government-funded research into detecting and monitoring psychological disorders among military personnel.
Now, Cogito is making a push into the commercial market, with a focus on selling the product to businesses with departments that field phone calls from customers. Today, the company is turning up the volume on those efforts with the announcement of a $5.5 million Series A funding round led by Romulus Capital, with participation from Salesforce Ventures.
“This is just a huge market opportunity,” Cogito co-founder and CEO Joshua Feast says.
Here’s how the product works: The audio recording of a customer service call gets routed to Cogito’s server while the conversation is taking place, and the company’s software analyzes things like pitch, tone of voice, pace of speech, and instances of overlapping voices. It then communicates insights to the customer service professional during the call with the customer. The software does this in two ways: it provides a score that assesses the quality of the interaction and a series of “targeted alerts around specific behaviors,” Feast says. These alerts might include a customer service rep coming across as tense, speaking rapidly, or interrupting the customer.
“The rep may have said absolutely nothing wrong, but maybe the customer is dissatisfied” because the tone of voice came across as impolite, Feast says.
The idea behind Cogito’s software is to instantly highlight possible trouble spots and give the employee a chance to recognize mistakes and adapt his or her communication style to the customer’s perceived preferences—while the call is still going on. “The goal is to help agents and customers have more caring and empathetic conversations and build rapport,” Feast says.
If the technology works well, it could lead to happier customers who are more likely to buy products from the company and recommend the business to friends. It could also make customer service employees more productive and engaged.
The problem is bigger than you might think. Cogito estimates more than three-fourths of customer interactions with businesses still take place on the phone. “America’s largest companies have wrung all the costs they can out of their customer care organizations,” Feast says. “And now they’re really viewing improving customer experience as one of their number-one strategic priorities.”
Meanwhile, there are high turnover rates in customer service departments, Feast says. “It’s a very tricky problem, and we think we can really improve the work life of phone-based agents by valuing and elevating their soft skills.”
Cogito software could help accomplish this by providing objective measurements of customer service workers’ performance, Feast says. “A lot of representatives are measured on subjective measures, and it’s hard to know if they’re doing a good job or not,” he says. “Knowing if you’re doing a good job or not is one of the biggest drivers of work satisfaction.”
Cogito was founded in 2007, based off of nearly a decade of MIT research into technology that can capture and interpret “honest signals”—involuntary cues in speech and behavior that help in understanding a person’s intent, the company says.
Before the Series A round, Cogito had raised $1.25 million in venture debt funding, $2 million in equity capital from angel investors, and more than $5 million in grants, Feast says.
Cogito initially focused on the behavioral health applications of its technology. It has formed partnerships with hospitals, the Department of Defense, the Defense Advanced Research Projects Agency (DARPA), and the National Institute of Mental Health.
Those efforts will continue, but Cogito is devoting fewer resources to the healthcare side of the business. “We see the big commercial part of the opportunity for the company as being in customer care and customer service,” Feast says.
That’s what attracted the interest of Romulus Capital, says managing partner Krishna Gupta. Cogito is the third Romulus investment in a startup from Pentland’s research group, following Ginger.io and Humanyze, he says. (Interestingly, Humanyze’s customers also include call centers.)
“The concept of behavioral signals, the entire palate of tonality and voice is very compelling; it’s very technically challenging to break down the signal,” Gupta says of what Cogito is attempting. “When you look at that hard technical problem and apply it to customer care and customer engagement, which really hits across verticals and business categories, we’re very excited about how big that can get.”
Other potential uses for Cogito’s software could include incorporating it somehow with voice-based virtual personal assistants, such as the iPhone’s Siri, Gupta says.
But that’s down the road. Right now, Cogito is plowing money into sales and marketing to add to its half-dozen customers using the software in their call centers, Feast says. The company will also grow its team of 31 employees by five or six people, he adds.