Customer call centers — we’ve all contacted them, to complain about an incorrect shipment or a non-working product, to report a stolen bank card or reschedule a missed flight. One company is aiming to make this experience less frustrating for customers and more efficient for companies by using artificial intelligence to analyze and then automate actions based on conversations.

Co-founders Umesh Sachdev and Ravi Saraogi began developing Uniphore 11 years ago in India, incubating their technology at IIT Madras, the country’s top engineering institute. Now, with more than 160 employees worldwide, Uniphore has set up its second headquarters in Palo Alto and expanded to locations in Singapore and Dallas Texas in addition to India. Clients include major insurance, banking, IT services and other multinational companies like NTT Data in Japan and BNP Paribas.

“We are sitting on a treasure trove of opportunity,” Sachdev says. “Customer service is a very large opportunity — it’s a $350 billion industry. We all face it every week, every month as consumers. And there’s a ton that AI and automation will do in the future to make that experience better.”

Uniphore’s main offering is auMina and currently is the primary revenue driver at the company. Sachdev describes it as artificial intelligence combined with automation capabilities that assists the call center representative during and right after a customer call. For example, you call your bank to talk about getting a new credit card. The call center representative puts you on hold to look for the right type of credit card for you, then puts you on hold again to retrieve your information and fill out a form for you.

“Every time the representative puts your call on hold is an opportunity for AI to come in,” Sachdev says. Instead of you being placed on hold, auMina presents the call center representative with information about the right credit card for you as soon as you mention the type of card you are seeking. It also fills out the form as the representative is talking to you. After the call ends, auMina summarizes the interaction, saving about three minutes per call that the representative would normally spend typing up this summary.

“This is AI that is now assisting the call center representatives,” Sachdev says. “It’s helping them do their job better. It’s cutting down the time it takes at doing these tasks and therefore saves money for the enterprise. And for you the consumer, it is a dramatically better experience. You’re not getting frustrated because your voice is heard.”

Co-founders Umesh Sachdev (left) and Ravi Saraogi.

Akeira, another product, is growing even more rapidly than auMina. Its application is for the quick questions that don’t need to be answered by a human — for example, a question about when a payment is due. Think of it as a virtual call center agent, Sachdev says, saving time for call center representatives and saving money for companies.

The third offering, amVoice, is still in the early stages of development. It focuses on security. Instead of a call center representative verifying a customer’s identity with the answer to a secret question, date of birth or other personal information that can be found with a simple Google search, a customer’s voice is his or her password.

“You have to now rethink and worry about security and privacy,” Sachdev says. “What we don’t realize is much like our fingerprint and our face, our voice is unique. And so we can do a biometric check on our voice. We are reimagining security and privacy in the new call center world.”

All three products were released in 2015, when the company received its first round of major funding. Today, Uniphore is expanding well beyond these three offerings while remaining completely focused in the call center market. The company is in a major growth phase and the goal is to reach $100 million in revenue in 2021. Uniphore recently raised a $51 million in funding and has plans to grow to 300 people, 200 of them in India, by the end of their fiscal year next March.

So far, Uniphore technology is used in eight countries, including the United States and India, but the company hasn’t landed clients in Europe, Australia and New Zealand. Sachdev says those destinations are goals in the company’s expansion strategy over the next 12 to 15 months. By the end of 2019, half of Uniphore’s revenue will come from the United States and the other half will come from Asia, including India.

Interestingly, the software “learns” language in much the same way we learn language from the time we are born, through listening. And like humans, Uniphore software must learn to pick up the emotional context of speech and is a process.That’s one aspect Uniphore continues to work on for the future. If AI can be taught to accurately recognize emotion, coupled with a customer’s stated intention, it could predict, with the help of other data such as a customer’s history, whether one customer is likely to pay a bill or not, regardless of the words he uses.

In the meantime, Uniphore has been refining its services in response to insights from customers and the benefits they are seeing in deployments. One, a major U.S. telecom company, told executives that sending a summary of a customer service phone call to both the customer and the call center representative would help communicate a sense of transparency to the customer. That will result in customers feeling they have been heard and their concerns are being addressed.

“When you’re calling the anonymous call center,” Sachdev says, “our hope is that nobody comes out with the feeling that they weren’t heard, that they weren’t attended to.”