Automatic Speech Recognition - ASR
Highest technology infrastructure
Round the clock
Automatic Speech Recognition (ASR) is a technology which allows interaction of a person with a computer through speech.
Why speech recognition?
- It allows the callers to make transactions in IVR through speech instead of dialing (DTMF) which also brings cost effectivity.
- It allows you to always provide high quality services without the variability which is based on knowledge and skills of customer representatives.
- In DTMF IVR, it allows receiving service within a much shorter time by eliminating the most common problem of customers’ getting lost among the menus.
- It allows providing service on 24/7 basis.
- It has a user friendly interface.
- There is no limitation imposed by touchtone menus.
- It leads new technologies and enables you to gain prestige in the eyes of your customers and competitors.
How do you authenticate a person who calls your Call Centre?
First two letters of your mother’s maiden name?
Your place of birth?
Your date of birth?
Your father’s name?
Your mobile phone number?
With VocalPassword, how many steps does the Speaker Authentication consist of?
The authentication process is realized in the following three steps:
Enrollment: When a customer calls for the first time, he/she is informed that, after going through the current security process, he/she would be able to quickly make the transactions without being required to remember any information such as a password, and then offered to create a vocal password. The customer, who is transferred to IVR, repeats the pre-determined passphrase three times to create a vocal password.
Verification: When a speaker, who passed through the enrollment process, enrolled and then left his/her voiceprint, re-calls the system; he/she is automatically recognized from his/her Claimed ID (e.g. Customer Number, TRID No), and is asked to indicate his/her passphrase.
The received voiceprint is submitted to the verification system, and this voice is compared with the vocal password. If it exceeds the threshold value which is determined as a result of this comparison, the “matched” result is received from the system, and the caller is deemed to be verified. After verification, the caller is authorized to perform the transactions.
Identification: The proccess of identifying a caller among the user group who previously left his/her voice print in the system. When the identification and verification functions are to be run on an IVR instead of a live agent, then the product to be preferred is Nuance VocalPassword.
Nuance FreeSpeech is a unique biometric speaker verification system that transparently verifies speakers in the background of a natural conversation. It is the most comprehensive, natural language speaker verification system available today, enabling customers to utilize the biometric power of voice to provide secure, efficient, and convenient access to contact centres.
It runs independent of content, language and accent. It eliminates authentication questions which cause loss of time for customers, and realizes authentication within seconds by using biometric voice characteristics. Moreover, it allows call centres to monitor all calls and detect fraudulent activities by creating a monitoring list mechanism.
Decrease in the number of misrouted calls
Reduction in the cost of human resources
More efficient service with less cost
Faster transactions, satisfied customers
Call Steering eliminates the complexity of contact centre navigation with a natural language solution that lets customers use their own words to describe why they’re calling.
In Call Steering application; when a customer calls the call centre, he/she is greeted by IVR saying “How can I help you?” The customer describes, through speech, which service he/she wishes to receive, and is transferred to the requested menu within a few seconds.
DTMF versus Call Steering Comparison
Zero-out rate (10-40%)
Misrouted call rate (10-35%)
Decreased customer satisfaction
Up to 40% decrease in the zero-out rate
Up to 50% decrease in the misrouted call rate
Improved customer satisfaction
- Enriched IVR menu can be used.
- Thanks to horizontal flow, rarely used options can also be easily reached.
- Significant decrease in the rate of misrouted calls.
- Significant decrease in the zero-out rate.
- Call durations are reduced on Pbx, IVR and operator side, and cost saving is ensured.
- Decrease in the number of hangups caused by being lost in IVR.
- As its natural result, it can be observed that the number of incoming calls decreases and customer satisfaction is improved.
- All new services, for which automation is possible, can be automated.
- More efficient and effective services can be provided with less resources.
- Decrease is observed in human resource costs.
- Customer representatives, who are the most valuable and costly item of call centres, are freed for more complex works.
Personalized Customer Experience
High Sound Quality
Text-To-Speech (TTS) is a technology of converting a written text automatically into human voice. With the use of this technology, any written text, whether it is a short text such as name & surname or a long text such as an article, is easily listened to and converted into a comprehensible human voice.
With Nuance Vocalizer, the Text-To-Speech (TTS) solution, you can enrich your existing IVR solution, and provide a more natural and personalized customer experience with a lower cost. Based on prolonged researches and specialty, Nuance Vocalizer is used in more than 40 languages globally. Vocalizer allows you to create dynamic prompts by using Text-To-Speech technology with your static prompts, and enables you to easily manage prompts of your entire application. Furthermore, it allows you to meet your requirements for urgent prompt changes quickly, easily and cost efficiently by eliminating high cost and timing problems arising from working with voice talents.
- Naturalness: The speech output, which is vocalized naturally, guarantees a good listening experience.
- Creating a flexible speech: Voice level and speech speed can be changed for more dynamic effects during execution.
- Direct phonetic input: It makes it suitable to read offline phonetic databases such as navigation map data.
- User glossaries: Rule sets can be determined in order to accurately read the words which can be pronounced differently in the natural flow of a language.
- Adding the prompts quickly and easily: The recorded prompts which are available in the system, as well as dynamic prompts created in line with instant requirements can quickly and easily be merged to be compatible with the natural flow of listening experience.
Examples of implementation in Turkey:
- In IVR systems; addressing the customers, who are recognized with their telephone number, with their name or name and surname.
- Reading the address for address confirmation
- Reading the daily stocks and stock market information
- In navigation systems, reading the names of locations, streets as well as commands
- Deployment of IVR promts which should be urgently recorded
- Reading newspapers over the Internet
Making process by talking
Authentication with VocalPassword
Virtual Assistant for Mobile Customer Service
Nina (Nuance Interactive Natural Assistant) provides a unique mobile customer service experience by converting smart phones into a customer service assistant with which you can communicate through speech. Nina transforms iPhone and Android mobile applications – which make very different transactions such as paying bills, making service changes or finding out a reply to a simple question- into robust tools through which customers can receive services quickly and easily without any assistance (by themselves).
What Can Nina Do?
Reliability: It authenticates you by using your unique voice print.
Transaction: It allows you to make faster transactions with far less steps by interpreting the conclusion it draws from the requirement you indicated through speech.
Target Achievement: With a single expression, it perceives the customer’s request properly, and transfers the customer to the most accurate location where he/she can receive a reply.
Unlike a programmed robot, Nina provides a far more natural interaction experience with its intuitive detections.
Nina recognizes you and your voice. It understands your request from your speech expression. By combining the power of speech, writing and touching, Nina makes mobile interactions such as logging in, finding out the application features or making transactions much easier, more efficient and enjoyable.
“Where is the nearest ATM?”
“Pay my all credit card debts from my bank account!”
“List the expenses over 100 TL which I made this month in my credit card!”
This virtual assistant application, which can be controlled with voice, allows you to easily make many of your transactions through speech such as banking and insurance transactions.
Unlike a one-way communication made with commands, it provides a communication experience in the form of a dialogue.
Nina is interactive. In order to provide a user experience which can be realized in a conversation atmosphere, it provides a completely non-complex, natural communication experience with its complex technology infrastructure.
For maximum flexibility, a ‘Multimodal’ interaction feature which allows you to make transactions through speech and dialing
Its multimodal user interface is indispensable for a high quality customer experience. Nina has been designed to allow users to act flexibly during a transaction with its speech, writing or touch modes, and to select the interaction type of their preference.