Speaker Verification
By 2007, 80% of all organizations will have reached the point where they will no longer be able to protect themselves with passwords alone, but will be forced to strengthen user authentication using alternative security methods. Companies should start using stronger authentication methods in high-risk areas… now [1].
A new report shows a dramatic shift in identity theft crime activity away from online to the more traditional phone-based fraud. Over the past 12 months, 8.1 million Americans were victimized by identity fraud, a crime amounting to $45 billion; and fraudster access through mail and phone-based incidents rose dramatically, from 3% of ID fraud in 2006 to 40% in 2007 [2].
Identity fraud and unauthorised access to accounts is on the increase:
- CIFAS reported over 40 000 cases in first six months of 2007.
- Cybercrime UK reported over 200 000 cases of online financial fraud committed where criminals impersonated their victims to obtain money, credit or even employment.
- In the UK, the loss in the post of Families Child tax credit information on two disks sparked a nationwide fear of a wave of perpetrated fraud.
- In the US, the TJX data breach was termed “the biggest card heist ever” by Gartner “where serious damage was done” with 45.7 million cards compromised.
A solution lies in the use of Voice Biometrics Technology. This technology can be applied to Speaker Verification where speech is captured from a caller and compared to a previously generated stored unique voiceprint, and a score is produced to say how closely the two are matched.
Contact us for a Fluency Assessment of Biometrics to find out how Speaker Verification can revolutionize your call center.
Key Features & Benefits
The combination of Speaker Verification to confirm who the speaker is, a speech recognition engine to determine what the caller said, and business logic to determine how to handle the caller, results in strong two-factor authentication.
Speaker Verification can either be:
- Text dependent – Where the caller is asked to say a specific phrase. The advantage of this approach is that performance can be better. However, the caller is required to ‘train’ the system by calling in from various environments and repeating the password. This can be a lengthy process which can confuse and irritate callers.
- Text independent – Where the caller can use any utterance. The advantage of this approach is that the process is transparent to the caller.
A Speaker Verification solution is key to knowing whether a person is who they claim to be, and is particularly elegant because it provides two-factor authentication in a single action – what someone said and the way they said it.
Speaker Verification has the following benefits:
- Eliminates weak security
- Works from any phone
- Is user-friendly
- Is cost-effective as it leverages existing web and IVR investments

