How can we help?

Talkdesk Identity FAQ



When does your caller’s name show up in the identity menu entry in Identity?

Only after the first validation is performed, i.e. the next time they call in after the enrollment is done.


What is “Risk” and how is it calculated?

“Risk” is an average value based on the voice authentication risk and the phone validation fraud scores. Those are analyzed to reach a risky score (from 0% to 100%), from which a chart is derived to represent criticality, i.e. Green = 0 - 49%; Yellow = +50%; Red = +70%.


How is the “Voice Authentication” status calculated?

The contact’s voice is first analyzed by Guardian Identity, which then produces an authentication risk score (from 0% to 100%), resulting in a “Successful”, or “Failed” status.


How is the “Phone Validation” status calculated?

Guardian analyzes several phone parameters, resulting in a ranked fraud score (from 0% to 100%). It then applies one of three possible statuses:  Safe = 0 - 74%; Suspicious = +75%; High Risk = +85%.



Voice Biometrics

What is Voice Biometric Authentication?

Voice Biometric Authentication (VB) is the technology that enables identity recognition based on voice characteristics. It achieves this by measuring the unique characteristics that the individual’s physical make-up (physiology) creates when voice any sound. It is also called speaker recognition or voice authentication.


What is the difference between Text-Dependent and Text-Independent Systems?

A text-dependent system is one where the customer is prompted to repeat the same pre-determined phrase as the one used during enrollment. This is commonly referred to as Active or Prompt voice biometrics.

In a text-independent mode, no specific text is expected – the customer is free to converse naturally and speak any words. This technology is typically used while passively processing conversations.

The choice between text-dependent or text-independent technologies is often driven by use cases and customer experience, rather than performance or security.


Will I still be recognizable if I have a cold, or if I am taking some kind of medication?

Yes. Having a cold, or being on medication, does not change the measurable variables in your voice.

Furthermore, because VB engines exploit information that relates to the shape of the full vocal tract (i.e. the physiological make-up of an individual) the effect of a cold (or medication) will be minimal.


Is it possible to fool a Voice Biometric system by mimicking a user’s voice?

When we  mimic someone else’s voice, we copy language mannerisms  — high-level language-production features. Voice biometric systems, however, exploit low-level features that relate to each speaker’s vocal tract. It’s easy to copy the way a person is talking (accent, mannerisms) but impossible to alter the way speech is produced (effect of the vocal tract).


Can a recording fool a Voice Biometric system?

Using a recording device to playback another person’s voice is known as a Replay Attack.

Voice biometric engines can identify recording devices, using a number of techniques including
detecting the absence of the highest and lowest frequencies which, though not audible to humans, are detectable to VB engines. Additionally, the process of replaying creates distortion to audio, and this is measurable by replay detection algorithms.


Can the system distinguish between twins?

Even though identical twins have the same genes but different physical development, their vocal tracts will vary enough for voice biometric systems to be able to discriminate between their voices.


Is privacy protected?

The Voice Biometric solution is built on a three-pillar “Privacy by Design” approach that places an emphasis on data privacy, data protection, and data security, and ensures compliance with all regulatory regimes.



Contact ID

How can the system identify the caller? 

When a caller enrolls, a differentiator is sent to identify the contact together with the recorded audio. This combination of unique identifier and audio is the caller’s voiceprint.


Can a four-digit pin be a unique identifier?

No, they are not unique identifiers since they are arbitrary numbers and several callers can end up setting the same pin.


How can I set up a unique identifier for callers?

You can choose to add an Input IVR component that will be used to collect the contact’s unique identifier (Contact ID), which will be matched with the voiceprint that will be collected.

If you choose not to set any Input IVR component to collect a unique identifier, the caller’s phone number can be used as the identifier.


What is a typical phrase for an effective Voice Biometric Authentication?

User-friendliness is key and consequently, a passphrase should be natural and easy to repeat by the user. If the phrase is too long or difficult to remember, the user would get frustrated. 


What are the pre-set passphrases?

In English, the two passphrases available are:

  • “Please authenticate me with my voice”.

In Spanish, the two passphrases available are:

  • “Por favor verifícame con mi voz”.
  • Por favor autentícame con mi voz.

In German, the two passphrases available are:

  • "Bitte verifiziere mich mit meiner Stimme".
  • "Bitte authentifiziere mich mit meiner Stimme".

In French, the two passphrases available are:

  • "Vérifiez mon identité grâce à ma voix".
  • "Authentifiez moi grâce à ma voix".

We will have five passphrases for each of the four available languages: English, Spanish, French, German.


How many passphrase repetitions are required to enroll a caller?

For a successful enrollment, the number of passphrase repetitions may vary between two to five times, depending on some factors (audio quality, passphrase wording, background noise, among others).


Enroll and Authentication

What happens to the caller's voiceprint if they opt out of the enrollment process?

When a caller requests to opt-out, their voiceprint is erased.


What happens if the enrollment process isn’t successful?

If the call drops or the user hangs up when the enrollment is still in progress, then the next time s(he) calls the enrollment process will continue, after the agent ends the call.

Will the agent know if the caller failed to authenticate?

Yes. If the caller is not able to successfully authenticate, the agent is warned within Conversations of the failure to authenticate.


Phone Validation

What does Phone Validation mean?

Phone number validation verifies phone numbers and identifies suspicious numbers in near real-time with phone number reputation scores. It identifies fraudulent users and invalid user information with known phone number risk scores.


Where do I see the calls’ Phone Validation?

In the Identity tab, inside Conversations, Phone Validation insights are displayed for every inbound and outbound call, irrespectively of voice authentication.

In the Identity App, the Phone Validation insights are displayed together with Voice Authentication.


Is Phone Validation fighting ANI Spoofing?

No. Phone Validation is not set to identify caller ID spoofing. Talkdesk will be working on leveraging the SHAKEN/STIR attestation to detect ANI Spoofing. 


Can I make Phone Validation without Voice Authentication?
Yes, when configuring a Studio flow, you can add the "Validate phone" component without adding any of the Enroll or Authenticate voice components.

Besides displaying the phone validation insights in the Identity and Conversations apps, this component can also be used for condition-based routing. See here how.


Understand insights

Why is a phone number considered suspicious or high-risk?

Risky phone numbers and those with recent abusive behavior are directly indicated in the results:

  • Fraud Scores >= 75% are suspicious and are likely to be involved in suspicious or dangerous activities, but not necessarily a fraudulent user.
  • Fraud Scores >=85% are high-risk users that are likely to engage in malicious behavior. Scores in this threshold indicate recent or excessive abuse and fit the profile of a typical risky user.

Why are invalid numbers marked as high-risk?

If the phone is identified as invalid or disconnected, that likely represents a fraudulent user with an inactive phone number.

What does “recent abuse” mean?

When a phone number has been associated with recent or ongoing fraud, then it is flagged as recent abuse.


What does VOIP and prepaid stand for?

VOIP indicates if this is a Voice Over Internet Protocol (VOIP) or digital phone number. Phone numbers marked as VOIP are inherently riskier than wireless/cellular and landline numbers. If the phone number is associated with a prepaid service plan, it will be marked as “prepaid” and it will also be considered higher risk.


What happens if the caller is successfully authenticated, but there are no phone validation results?

If the contact unique identifier is not the caller’s phone number but rather an Account ID # or SSN #, or other, then there will be no phone validation results displayed.

All Articles ""
Please sign in to submit a request.