Glossary

Abandon Rate

Accents /Dialect   

Dialect is the form of language spoken by people in a particular region or group and accent is the way you speak a language in a particular style. Speech recognition applications are designed to recognise every type of accent and dialect to ensure accuracy.

Accuracy                                  

Typically used in speech recognition when referring to whether a speech application has understood what a caller has said. Many factors affect accuracy such as the performance of the underlying speech recogniser, the grammar, the prompt that the system speaks to trigger the caller utterance, and ambient noise. It is the task of the speech application provider to take these into account when designing a robust and user-friendly speech application.

Ambient Noise

The background sound and level at a given location. The higher the ambient noise, the more difficulty a caller may experience when trying to use a speech application.

Applications

The services that a speech solution can provide for example; call routing, real-time information, bill and account payment and product purchase.

Automatic Call Distributor

(ACD)

Automated Speech Recognition

(ASR)

See Speech Recognition

Automation

Capacity

For an IVR or Speech Service, capacity refers to the number of simultaneous calls that a service can handle.

Caller Line Identification

(CLI)

A service that transmits a caller's number to the called party's telephone equipment. The number may be displayed on the called party's telephone equipment, or can be used in applications for example to identify repeat callers.

Confidence

Speech recognizers provide an ordered list of possible things a caller may have said, each of which has number defining how sure the system is that the caller actually said that utterance.

Concurrent Calls

Cost to Serve

Dialog

A dialog refers to the naturally spoken words, phrases and sentences that make up the conversational roadmap of the voice application. A dialog defines what callers can say and what they will hear at any given stage of the interface.

Dual Tone Multi Frequency

(DTMF)

The tones produced by pressing keys on a telephone. DTMF, also called Touch-Tone, is often used as a way of sending data to IVRs.

Grammar

A file that contains a list of words and phrases to be recognised by a speech application.

Homophone

Hosted Solution

Hosting Centres

Identification and Verification

(ID&V)

Identification and verification is typically used to refer to the process of identifying someone when they call in to a contact centre, and verifying that information. E.g Account number and PIN, or postcode and date of birth.

Intelligent Call Routing

Software that identifies calls based on network provided, call entered, or customer supplied data and then directs them to a specific location.

IVR - (Interactive Voice Recognition or Intelligent Voice Recognition)

A system that allows callers to interact with a computer, using a telephone. An IVR may use speech recognition, DTMF, or a combination of the two.

Lexicon

Liveness Testing

The goal of liveness testing is to determine if the biometric being captured is an actual measurement, from the authorised, live person who is present.

Managed Service

Allows clients to focus on core business operations through a company such as Eckoh managing the service needs and improving efficiencies while reducing cost.

Natural Language

An approach to speech application design that encourages users to speak naturally to the system.

Payment Card Industry Data Security Standard

(PCI DSS)

PCI DSS is the payment card industry requirement for all organisations that store, process or transmit credit or debit cardholder data. Eckoh is a level 1 service provider which is the highest standard possible.

Phonemes

The basic unit of sound. In the same way that written words are composed of letters, a spoken word is composed of various phonemes, though they may not line up precisely.

Prompts

A prompt is the dialog played to a caller either to ask a question or to provide feedback. A prompt can be played from a pre-recorded file, generated from text , or a combination of the two.

Public Switched Telephone Network

(PSTN)

PSTN refers to the international telephone system based on copper wires carrying analog voice data.

Scripts

VoiceXML code designers use scripts to quickly and accurately define the dialog, the prompts to be played and what grammars are active (i.e., what the caller may say ay any given point in the dialog).

Self-Service

An application that allows a caller or customer to conduct a transaction or enquiry without the assistance of a Customer Service Representative or Agent.

Short Message Service

(SMS)

SMS comprises standardised communications protocols that allow the exchange of short text messages between fixed line or mobile phone devices. SMS can be used in self-service applications either as a request/response service, or simply as a confirmation of a transaction.

Speaker Verification

Identifies a caller based on their vocal patterns. This feature enables a caller to identify themselves by voice rather than by entering a password for security purposes.

Speech Recognition

The ability of a program to identify words and phrases in spoken language and allow applications to understand what a caller has said and act upon it.

Synonym

Different words with almost identical or similar meanings, for example buy and purchase.

Text to Speech

(TTS)

The process of converting electronic text to spoken words. Text to Speech software converts text into audio output, enabling large quantities of text-based information to be heard over the telephone.

Two-Factor Authentication

Using any two methods of authentication together to process an identity claim, for example; something you know (Postcode) and something you are (Voiceprint).

Utterance

Typically used in speech recognition when analysing what the caller has said. An utterance refers to a unit of speech bound by silence.

Voice Biometrics

Identifying people by voice, or voice print by measuring unique biological characteristics for the purpose of comparing unknown samples against known samples, usually with the goal of confirming someone's identity.

Voiceprints

A voiceprint is a set of measurable characteristics of a human voice that uniquely identifies an individual voice. Similar to a fingerprint where the various ridges, troughs and curves on a finger provide a distinct method of identification, the differing physical components of a human throat and mouth produce a unique sound that can be analysed, measured and the results securely stored. The stored data is known as the VoicePrint. Note that recording of the voice sample is not stored.

To verify an identity, a second sample of speech is analysed and the measurement process is repeated. If the result of the calculation matches the result previously obtained then the identity can be verified.

Voice User Interface

(VUI)

VoiceXML

(VXML)

Voice Over Internet Protocol

(VOIP)