This Call May Be Monitored or Recorded for Quality Assurance Purposes

We’ve all heard this phrase when trying to reach an account representative or customer service department. Understandably, most institutions use recorded conversations for training purposes and to improve customer satisfaction. However, advancements in voice analytics will soon give new meaning to the phrase.

The scenario is all too familiar:  You have an issue with a bank, wireless or Internet provider and need to speak with somebody, so you call a 1-800 number to connect with a representative. First, you provide your account number, then your date of birth. Next, you’re prompted for the last four digits of your Social Security number, and then (if you’re “lucky”), you get to re-enter the same information two or three more times before you finally get to talk to someone. Just before you’re connected with a representative, you hear, “This call may be monitored or recorded for quality assurance purposes.” The process may be painful, but it’s necessary to protect clients against losses resulting from identity theft.

According to the Social Security Administration, identity theft is one of the fastest-growing crimes in America. A dishonest person who has access to your personal information, such as date of birth, Social Security number and a few personal details, can use it to open new accounts and stick you with the bills. Alternatively, a fraudster could gain access to your bank accounts or IRS refund and pocket large sums of money without anyone knowing. In 2014, the IRS paid out an estimated $5.8 billion in fraudulent refunds related to identity theft. That number is expected to hit $21 billion by 2016. Multiple agencies are working to stop and prosecute identity theft. The latest developments in voice analytics may provide a solution that will greatly assist these agencies in their efforts.

Over the last few years, a new generation of voice analytics technology has emerged through a number of high-tech vendors. The new voice analytics solutions provide the ability to analyze every call in real time. Voice analytics can be used to identify emotions such as anger, distress or frustration. Any phrase, keyword or combination of words can be used to identify red flags. In a more sophisticated form, real-time voice analytics can be used for biometric authentication (voice biometrics), which can prevent imposters from using stolen personal information.

Several large banks in Europe, Asia and the U.S. have quietly launched voice recognition services that authenticate customers by the pitch, tone and modulation of their speech patterns. With voice recognition technology, customers no longer need to enter their card number, PIN and security answers to authenticate themselves. Their voices act as the password for banking transactions through the call center.

This “voiceprinting” technology could soon be adopted by business and governments alike. Potential applications are limitless and might be used to:

  • Initiate and authorize fund transfers between investment accounts
  • Confirm disbursal amounts to vendors
  • Authenticate taxpayers for refunds
  • Verify the continued existence of pensioners or those receiving paychecks
  • Distinguish between human and nonhuman behavior patterns, e., robotic or BOT activity
  • Protect personal assets through cell phone applications like Apple’s Siri

So the next time you call your bank and hear a friendly voice saying, “This call may be monitored or recorded,” that may be the last time you hear those words before they’re replaced by voice biometrics.


Michal has more than eight years of experience providing litigation support, forensic accounting, audit and advisory services in a variety of industries, including banking, construction, financial services, governmental, higher education, manufacturing, not-for-profit, real estate, retail, social services and transportation.

Michal Ploskonka – who has written posts on BKD Forensics.

Leave a Reply

Your email address will not be published. Required fields are marked *