Pindrop

History

The origin of Pindrop was a concept developed at the Georgia Institute of Technology. Vijay Balasubramaniyan, a computer science graduate from India studying the traits of wanted versus unwanted phone calls, teamed with his thesis advisor Mustaque Ahamad to launch a VentureLab project called Telineage in September 2010. Balasubramaniyan, Ahamad and two colleagues presented an ACM paper on what would become Pindrop's core technology.

Balasubramaniyan's idea was to acoustically fingerprint phone calls and associate that data with the phone number. In 2011, he completed his studies, partnered with Paul Judge, a fellow GT alum and security industry veteran, and founded Pindrop as a voice security company that combats fraud by analyzing and assigning risk to phone calls.

The company obtained a license to the intellectual property from Georgia Tech Research Corporation, while VentureLab supported the company for a Phase I commercialization grant from the Georgia Research Alliance. In the same year, Pindrop placed second in Startup Riot, was named one of Top 40 Most Innovative companies in Georgia by the Technology Association of Georgia (TAG), and won the TAG/GRA Business Launch Competition. It also won a National Science Foundation SBIR award.

In 2012, Pindrop raised $1 million in funding from Andreessen Horowitz, other venture capital firms and several angel investors. In June 2013, the company closed an $11 million Series A investment round led by Andreessen Horowitz and Citi Ventures and also including Felicis Ventures and Redpoint, using the funds to scale up engineering, operations, sales and marketing in the US, Canada and Europe. In February 2015, Pindrop raised $35 million in a Series B round led by Institutional Venture Partners.

Their employee headcount totaled 70 at the beginning of 2015, which increased to 100 by June. In July, Pindrop released Fraud Detection System 2.4 (FDS), providing new tools for fraud analysts and call center technologists. Pindrop's 2015 revenue tripled, while its customer base doubled.

Pindrop raised another $75 million in 2016. Google Capital led the Series C round, with participation from Andreessen Horowitz, GV and others, bringing the company's total funding to $122 million. It was a rare case of both Google Capital and Google Ventures (now GV) investing in the same startup.

Pindrop claims that its solutions, supported by its patented Phoneprinting technology, are in use by a number of insurance companies, government agencies, retailers, and three of the four largest banks in the United States. In total, the company has screened more than 360 million calls. The company declined to name the clients, however, because of its nondisclosure agreements. The company plans to add 100 new jobs in Atlanta and another 50 at its locations in San Francisco and in the UK, where it already has a 10-person office. It will lease 45,000 square feet in Tech Square and plans to expand its presence in the EMEA region. It is setting up operations in Latin America and planning to move to East Asia later in 2016.

Services

Pindrop helps protect against fraud, including those using robocalls and faking a caller ID from anywhere in the world. In 2015, Federal Trade Commission singled out robocalling as the No. 1 consumer complaint in the U.S., with about 170,000 complaints a month. Pindrop's publicized report The State of Phone Fraud 2014-2015: a Global, Cross-Industry Threat found that 86 million calls per month in the U.S. are phone scams. It also found that 1 in 6 phone numbers calling a consumer is a robocaller and 2.5 percent of U.S. phones receive at least one robocall per week. Finally, it revealed that one in 2,600 calls is fraudulent, meaning that a large company can receive more than 30,000 phone fraud attempts each year, or 80 each day. Such calls create the loss of more than $20 billion annually in the US alone.

Phoneprint

Pindrop's "acoustic fingerprinting" technology integrates with companies internal systems and identifies people's voices, locations, and devices. This is added to a database for future reference and to help separate legitimate callers from scammers.

The company listens for 147 different features that help identify the uniqueness of a device and attaches it to a caller. To create metadata, Pindrop analyzed millions of phone calls in telecom databases from around the world and used machine learning to turn that information into usable content. By analyzing both the audio of a call and the metadata it has about a caller, the phoneprint reveals whether the caller is using a cell, landline, or VoIP phone; where the call really is coming from; and whether the caller has been seen before. It looks for evidence of frequency filters and codec artifacts, for example, and analyzes the calls for packet loss and dropped frames. In packet loss, "pindrop"-sized bits of audio drop out, which is where the company's name came from.

Based on the analysis, the service generates a risk profile and a Pindrop score for each call. Any call receiving a score of 60 or higher suggests that person may be using caller ID spoofing or voice distortion technology. Analyzing millions of samples from call centers, Pindrop can identify specific criminal groups. An example is a Nigerian ring nicknamed "West Africa One," with a dozen members according to Pindrop, which has identified the skill levels of each member.

Other tools

Pindrop has additional tools that update the company's fraudulent call database with relevant information. Phoneypot is a telephone honeypot with about a quarter-million phone numbers that are not being used by real people, which Pindrop uses for research. Workers enter the numbers into sweepstakes and online databases to collect data from millions of calls from robo-callers, debt collectors, and telemarketers. Among trends, the researchers found that older phone numbers attracted more calls than newer ones.

Topic Modeler is a proprietary online complaint collection tool that aggregates data on suspicious numbers from complaint sites, online communities, and web forums. Fraud Detection System (FDS) combines the company's Phoneprint technology and voice biometrics capabilities with an analytics-driven, unified workspace to more quickly detect and curb phone fraud. It allows fraud analysts to collect call samples, playback and annotate calls, query the system and investigate possible patterns of fraud, all in one platform.

Awards and recognitions

Cybersecurity Ventures ranked Pindrop at #72 among the hottest 500 cybersecurity companies to watch in 2016. InformationWeek included it in the top 100 cyber security companies to watch in 2016. Finally, Dark Reading named Pindrop in the 20 cybersecurity startups to watch in 2016.

SINET chose Pindrop as one of the top 16 emerging cybersecurity companies in 2015 from a pool of 122 applicants from around the world. The company received the 2015 Innovation Award from the Association of University Research Parks.

Ahamad and other researchers presented a paper on Pindrop's database PhoneyPot at the Internet Society's 2015 Network and Distributed System Security Symposium. The paper, outlining how to construct a successful telephony honeypot, won a distinguished paper award.

SI 100, an annual listing by Silicon India of the top 100 technology companies founded and managed by Indians in the U.S., included Pindrop among the top 10 enterprise security companies in 2013. MIT Technology Review placed Pindrop CEO Vijay Balasubramaniyan among 35 Innovators Under 35 in the Inventors category. At the 2013 SC Magazine Awards, Pindrop won the Rookie Security Company of the Year award, while its Fraud Detection System won the Best Emerging Technology award.

Pindrop was named one of the 10 Most Innovative Companies at RSA 2012. Gartner included the company among the Cool Vendors in Enterprise Unified Communications and Network Services in 2012.