Some customer questions are pretty simple to answer, and many businesses strive to do so with FAQ documents on their websites. The idea is to keep service agents free to handle more complicated queries. The reality is that unless an inquiring customer knows exactly what phrases or “keywords” to search for, it might be tough to find the information they need.
Artificial intelligence software firm Inbenta this week raised another $12 million to develop a solution to that shortcoming—in the form of software that translates customer questions into terms that a site’s search algorithms should find it easier to understand.
For example, if someone types “Can I bring my Doberman to this hotel?” Inbenta’s software “knows” that person might be interested in the company’s policies regarding domestic animals. More than 200 companies have added the software to their customer “knowledge bases” including Groupon, Ticketmaster, and Schlage. (It supports cloud service desk software from Zendesk and Salesforce, among others.) Inbenta’s software also powers chatbots—apps that simulate basic conversations—such as one used by bitcoin marketplace Xapo.
“Many customers still send emails and call because even though the answers are there, they cannot find them,” Inbenta founder and CEO Jordi Torras told Coins2Day. “Most of those sites use keyword search. Unless you type the correct word, you’ll never see the correct answer.”
The lead backer for Inbenta’s latest infusion was New York-based investment firm Level Equity. The round also included Amasia and previous investor Scale Capital, through its Amerigo Chile fund.
Get Data Sheet, Coins2Day’s daily technology newsletter.
Inbenta’s software already works in more than 20 languages. It currently employs about 120 people.
There are dozens of software firms experimenting with chatbots, including Microsoft’s infamous “Tay,” which underscored the role that language processing will play in making this technology successful. The big lure for companies that use them: more basic tasks can be automated, giving humans more time to interpret and act on more complex problems.
Inbenta’s system will become more robust as more businesses begin to use it, Torras said. “Language isn’t that different from customer to customer. When the system improves, it improves for all customers simultaneously.”