Startup dives into payments data to modernize customer service

Register now

Customer support may seem like it's innovating through the use of artificial intelligence and chatbots, but all of that is built on a foundation that's stuck in the stone age.

By innovating on top of a rusting engine, companies miss a big opportunity to grab a trove of new information to make customer interactions less annoying and more proactive, according to tech entrepreneur Brad Birnbaum.

"A decade ago there were no smartphones or message bots, so a lot of the tech behind customer support is from another generation," said Birnbaum, one of the founders of Kustomer, a 2-year-old company that pulls payment data, shopping analysis and other information to add more depth to traditional CRM.

In most cases regarding customer support, retailers don't have access to payment gateways, back-end systems and other customer data—and are only able to access to data that is tied to a specific ticket.

"You don't want to treat people like a 'ticket,' but as people who have other activities and needs besides the support call," Birnbaum said.

As a consumer engages support, Kustomer sifts through past interactions cross different channels, such as online, chatbots and phone, and then applies machine learning and a connection to the retailer's back-end system.

This creates myriad opportunities, such as allowing most of a business to handle customer service instead of handing a consumer off to a specific department. The user also has access to data outside of the service ticket, so information on sentiment analysis, social networking tied to the company, a customer's profitability and problems with other products at specific dates and times come up, along with a series of past transactions and the channel used for payment.

The idea is to move beyond a conversation that's about something "not working" to an informed session, in which a resolution can be obtained and acted on. Kustomer connects the service ticket to the payment gateway to enable refunds and/or accept payments for cross-selling or upselling.

A new product called Campaigns is more proactive, reaching out to consumers who abandoned a shopping cart and, based on access to data outside of that particular interaction, determines if it's a low-value or high-value customer. The agent can reach out with intelligence-driven suggestions to resurrect the abandoned shopping session, expand it, or suggest other products. A workflow is created that communicates with the payment gateway to process the transaction—without having to sign in to another program, Birnbaum said.

Earlier in their careers, Birnbaum and Jeremy Suriel, another Kustomer co-founder, founded Assistly, a cloud-based customer management system that they sold to Salesforce for $80 million in 2011. Salesforce rebranded the product as

Birnbaum and Sureil stayed at for about three years, then worked at Airtime, a video chat startup.

They have put elements of that experience to work at Kustomer, with the idea that chat, AI and modern social networking can be applied to customer service. "We saw a lot of point solutions for customer support, but what we realized is companies want to know more about customers," Birnbaum said.

In a short time, Kustomer has signed up clients such as, Bespoke, Priority Bicycles and Ring, which was just sold to Amazon, giving Kustomer a deeper reach into e-commerce. It has also drawn more than $12.5 million in funding from Canaan Partners and other investors.

The application of machine learning customer service opens up ways to move beyond direct service, bringing in elements such as easy payments and marketing.

"With more data and insight into things such as customer sentiment, personality, likes and dislikes, bots may be able to read human emotion better than a human," said Tiffani Montez, a senior analyst at Aite Group. "A couple of ways being considered to help humanize a bot transaction include scoring customer sentiment and using that to help steer the conversation and handing it off to a live agent when needed."

For reprint and licensing requests for this article, click here.
Artificial intelligence Customer data Payment processing Venture funding