Hitmetrix - User behavior analytics & recording

DialogTech uses AI to screen out spam

The spam phone call is a familiar scourge to the average person. It arrives as an interruption, with an unrecognizable phone number appearing on the screen of the smart phone. 

It’s bad for the individual. It’s worse for a corporate sales department. They can quantify the waste of time in dollars. Looking to save clients a few bucks, (okay, maybe tens of thousands of bucks), DialogTech today introduced SpamSentry, an automated software module that will screen out spam phone calls as part of its Voice360 platform.

SpamSentry relies on machine learning, where the software can actively detect the nature of the phone call and screen out the spam, at some cost savings for the client. Increasingly, many companies are approaching customers through their smartphones, explained Irv Shapiro, CEO. These firms will make that approach by drawing from a pool of phone numbers, and building up a customer profile when one of those unique phone numbers is used to call about a product or service. 

Spammers cycling through those 10-digit queries are slowly polluting that pool of unique phone numbers, “invalidating the attributes you’ve been trying to get,” Shapiro explained. This becomes the phone equivalent of click fraud, he contended. It gets worse if spammers are dialing service desks as part of their troll. Those calls will end up in hold queues alongside legitimate calls, lengthening wait times for customers. 

“There is a direct charge for every one of those calls,” Shapiro said. “Misattribution of live calls can be perceived to be a conversion when in fact they are not,” added Steve Griffiths, senior vice president of marketing, strategy and 
analytics. The expense of marketing goes up because decisions are being made based on flawed data, he said. 

SpamSentry was born because of this necessity. “We spotted the problem because of customer complaints,” Shapiro said. DialogTech’s R&D department was tasked with finding a solution. The unit had to find a way to distinguish between a genuine customer call and a spam call. This came down to relying on a neural network to pick out a pattern. But to help build that benchmark, DialogTech staged its first “Spam Olympics”. About 30 to 50 
employees screened tens of thousands of phone calls via computer screen, clicking on “spam” or “not spam” to label each one of them. While the winners of the competition got special prizes, the company got something more valuable: the basis for an algorithm that is more than 99 percent accurate when screening incoming phone calls. The result is that SpamSentry can recognize the pattern of a spam phone call and screen it out. 

“The beauty of the algorithm is that it is self-learning,” Griffiths said. It will adapt to the changing nature of spam calls, which is expected as spammers will change their technique and approach to try to maintain their effectiveness. 

SpamSentry becomes part of DialogTech’s Voice 360 platform, effective today.

Related Posts