Inbound Call Attendance

Inbound call attendance was the first area in which IVRs have traditionally been deployed. Speech technologies have proven their value to improve the way this service is performed in enterprises and call centres:

• Dynamic welcome messages: in addition to the traditional standard welcome message, companies may like to broadcast dynamic general or personalized information or alert messages. Using text-to-speech, the content of these messages may easily be updated on-line by enterprise business users, in case of exceptional situations.

• Enterprise virtual receptionist: through a company voice directory, the caller may reach a person or a department, just by saying its name, the company’s department, or even the type of service or product involved. Call filtering can also be introduced: the called party can accept or reject a call or direct it to a voice mail. In each case, the system manages the relation of the dialogue with the caller.

• Call attendance and routing in call centres: manages the first level of contact with the calling customer. Enables to qualify the purpose of the call, which will then be handled either by an automated speech-enabled module, or by an agent. Speech recognition allows a much larger amount of potential call reasons to be managed, with a higher comfort for the caller.

• Wait time management: in case the inbound call could not be immediately transferred to the desired person, the position in the ACD queue and average waiting time before being served may be announced to the caller. This allows callers to manage their call in a more flexible way and therefore gives a dynamic
image of the company. During waiting time, dynamic flash information messages
may again be played.

• Call back requests: when the requested service or person is not available, a request for call back may be registered together with the number to dial, the expected time range, and an optional voice message.

• Voice-Mail: if the caller wants to leave a voice message to a recipient, it may be recorded by a speech application. This information may be sent by e-mail to the desired recipient, together with an optional SMS notification. In addition, the speech application may store the voice messages in a database and provide a web interface to manage its content.

• Caller identification: allows to identify the customer or prospect either automatically from his calling number or from information that he/she provides (customer or contract ID, dial number, zip code, address, names, etc…), which is then searched into company CRM databases or public directories. The preliminary identification of the caller allows the dialogue and menu options to be further
personalized.

• Call transfer with IVR attached context through CTI: when the call is transferred to the desired person or department, dynamic contextual information is passed to the recipient desktop, in order to allow a popup display of the customer profile and a summary of the information collected by the IVR.

• Conference calls: thanks to the CCXML standard, a standard to manage call control in VoiceXML environment, not only can a speech application be used to attend and transfer a call, it may also organize a conference call between several persons such as an inbound caller and several contact persons in the company (ex: a call centre agent + a back-office expert).

• Video Conference: the latest evolutions of the VoiceXML platforms now allow video information to be streamed in addition to audio information. As a result, inbound callers using a third generation mobile phone device may enter a videoconference with the contacted person or call centre agent.

Outbound Dialling

Outbound dialing is an application automating the dialing of internal as well as outgoing calls. In the first case, the general phonebook of the company is use; in the second case, the system identifies the caller and loads the corresponding personal phonebook. When employees call, they are asked to say the name of the colleague or correspondent they want to talk to; the system dials the corresponding number, and connects both parties. Moreover, each employee from a set of different numbers, being their office extension, mobile phone or even private home phone, can access this functionality.

Directory Assistance

As for the module explained above, for very large directories more information than just the name of the person is needed to do a successful search. Often the city name or street name is asked first, before the name of the person is pronounced.

Location Finder

Large companies and administrations often operate a network of points of sales and services like shops, stores, dealers, offices, agencies, brokers, hotels, ATM terminals, garages, petrol stations, service centres, night duties, pharmacies, hospitals, etc. Similarly, tourist offices publish information about tourist attractions, cultural and sports event venues.

As a result, call centres are often asked to provide customers with routine information about those locations, contact information, opening schedules, and types of products or services available. Speech technologies now enable the set-up and rollout of location finders that operate 24/24 in automatic self-service mode, which improves customer satisfaction and reduces call centre operating costs.

A speech-enabled location finder allows callers to get information about a Point Of Interest (POI) they want to visit or contact. A POI can be retrieved through multiple search criteria such as location (country, region, city, or a maximum distance from a specific place), the available products or services, or the opening schedule.

A speech-enabled location finder may include a geographical database covering the target areas with a lexicon of the right phonetic pronunciations of the city and street proper names, as well as the hierarchical organization of locations (country - regions - cities – districts).

Marketing Portals

Enterprises and administrations usually have a catalogue of products and/or services. Communication about products and services is a fundamental part of the organization’s marketing plan. Up-to-date information about deliverable products and services can be found on a web site or obtained from a call centre. However, handling those calls for information is a routine job that does not generate any immediate value for the enterprise. This pleads for the automation of this service.

A voice marketing portal accesses the company product database that describes product and services in terms of reference, name, category, features, detailed description, target public, price, availability and delivery. Thanks to the integration of speech recognition, callers can express their search criteria at any time throughout the dialogue process, by telling the system about the name and/or features of the requested products. Product proposals and descriptions are provided dynamically with text-to-speech technology, so that the voice portal always publishes the latest up-to-date information about products in the company database.

First-time callers who don't know how to browse the catalogue by voice can be guided by automated prompts that present the product database contents.

Order taking

Retail companies have implemented plenty of self-service e-business applications over the last years, which enable customers to order products and services on-line over the web.

More particularly, home shopping retail companies, in many sectors like textiles and clothing, cosmetics, food and beverages, flowers, office supplies, sports, culture, leisure, home appliances, electronics, banking and insurance, travel agents, etc., are used to interacting with their customers through call centres.

The VoiceXML standard now makes it possible to implement self-service voice applications for product and service ordering that leverage the enterprise web-based e-business infrastructure. Most recent speech technologies (TTS and ASR) may so be used in seamless integration with the back-office Enterprise Resource Planning (ERP) and Customer Relationship Management (CRM) applications. Typical services provided cover customer identification, selection of order items, options and quantities, and delivery specification and payment.

Order Tracking

Directly related to on-line ordering, the order tracking applications may even more benefit from speech technologies. Although an increasing part of customers are using graphical web interfaces to order products, they still mostly use the telephone indeed to claim about the status of a product or service order.

After due customer identification, the speech application may query the enterprise CRM application or database and provide the caller with real-time information about the current order status (content, status, expected delivery date,)

Agenda management

This module is often used to automate appointments, or to automatically fill in the agenda of a doctor or sales representative, for example.

There are several ways of doing this; these are only a few examples.

1) The system can read to the caller the first available date and time for a next appointment. This can be read back by the computer, using Text-to-Speech software.

2) The caller can than say something like: “ I accept”, or “Yes”, or another confirmation, if he accepts this appointment.

3) If he does not accept this proposed appointment, he can ask for “the next available date”, or propose a “date in the second half of August”.

Depending on how the agenda of a particular service is constructed, or depending on how the dialogue is structured, or the way the agenda is filled up, the dialogue with the Agenda Management can be customized, or structured upon request of the involved parties.

Caller identification and speaker verification

Next to traditional ways to identify the caller (he can be asked for his name, address, etc., or the system can identify the calling number), speaker verification is now available.
 

During a one-time enrollment phase, callers provide some speech samples over the phone. The speaker verification software creates a voiceprint and stores it. This voiceprint models features of the speaker’s voice, but it is not a recording of speech; in fact no audio samples are stored.

During verification, callers pronounce the requested information such as name or ID number. The speaker verification software compares these live speech samples to the stored voiceprint. Voice authentication tracks the physical characteristics of more than just the vocal cords, so registered callers with a slight cold will not have any problem getting verified. Only extremely bad colds may have an effect on verification.  Potential imposters trying to mimic other people’s voices still have a different voice, so their voiceprint will not match. In this way secured access to sensitive information is not compromised.

Studies and live deployments have shown that when customers are provided the option to use voice authentication, they will use it! In the US, over 150 million secure, automated calls are being processed  annually through voice verification applications in the following domains:

• Account management of financial, benefits or healthcare information
• Field service automation
• Sales force automation
• Corporate dialers
• PIN reset
• Facilities access

Customer care

Customer care is a strategic part of the overall customer relationship, where customers get the opportunity to access and exchange their personal information. Typically, access to personalized information happens through web sites or through calls with live operators.

A growing number of contact centres offer their customers a self-service round the clock personalized access to information. After successful identification, callers may access a wide range of information, such as administrative data (contact information, list of subscribed services, applicable rates, etc.), order status and shipment tracking, consumption reporting, current account status, accumulated fidelity points.

Speech-enabled applications may easily extract this information on-line from the back-office CRM enterprise systems. Text-to-speech technology enables the presentation of these dynamic data in their current status.

If the caller is expected to visit the voice portal frequently, the voice application may save the used search criteria in order to be reused in subsequent calls.

At any time, callers may optionally be transferred to a live operator on simple voice request.

Customer registration (pre-paid card registration)

A very heavy burden on contact centres is linked to the registration process of customers. The legal requirements for mobile telco providers are evolving toward the mandatory registration of the owner of pre-paid card prior to activation. This requirement is already effective in several countries outside Europe and knowing that this may represent a significant increase of customer interactions, some operators have chosen to use speech self-service application for the registration of their customers.

The process can handle the transaction completely without manual intervention and proves to be extremely effective and efficient.

Helpdesk

Resetting expired, forgotten or compromised passwords is a routine part of life for millions of information workers. It’s also a tedious, time-consuming process that complicates already busy schedules, cuts into worker productivity and costs companies millions of dollars per year. Password resets are the second most common reason workers call help desks, accounting for about one in four help desk requests. Other examples are frequently asked questions, or problem reporting and tracking.

Speech applications (with speaker verification) that enable users to reset their own password over the phone, cut the cost per reset and reduce time spent by IT help desk staff on password reset tasks by a minimum of 30 percent. Both local and remote users can quickly and easily reset passwords with a simple phone call.

Phone Banking

By providing real-time account access via speech recognition technology, customers can simply ask the system for information such as, “what’s my account balance?” Customers can also perform financial transactions such as, “transfer two hundred euros from checking to savings”. Customers are in control and will not feel limited by traditional telephone menus.

Speech shortcuts and support for natural language phrases allow callers to stay in control of the system. Touch-tone is also available as a fall back option for touch-tone power users.

Speech is used to acquire account details, account history, funds transfer, stop payment, branch hours (holidays), general information, after hours’ messages, branch and ATM locator and transfer to operator.

One of the first speech-enabled applications in the banking world was developed to trade stocks directly from the phone. Identified callers have the choice to listen to actual share prices, or give sell/buy orders on their own portfolio.

Banks offer these new products to their clients to differentiate their e-business self-service strategy.

Travel information & Booking

When traveling, access to up-to-date information can make the difference between catching a plane and sleeping in an airport hallway. About a decade ago, airport information lines were among the first IVR applications to be speech-enabled in the Unites States. Asking for arrival or departure times by simply saying the name of a city or airport is so much easier than having to type in a partly alphanumeric flight number or some other arcane code.

Likewise, knowing the scheduled departure or arrival time of a bus or train before getting to the station makes for a much more relaxed travel experience. Offering this extra phone service on top of existing channels like the real-time information boards is perfectly feasible. The information is already electronically available, so why not exploit it, without overloading the call center with routine questions?

Other typical traveler’s information needs include (static) timetables, ticket bookings, lost luggage, car rentals, hotel shuttles, etc. All of these involve names rather than just numeric codes, and are therefore good candidates for speech-driven automation.

Real-time traffic information

The evolution of the economy towards globalisation dramatically increases our need for mobility, leading to an increasing usage of various transportation means. Growing traffic congestion, maintenance works, severe weather conditions, strikes in public transportation often lead to huge loss of times. In such context, travelling people need accurate real-time information to enable them to adapt their travelling behaviour to current conditions and take appropriate decisions to optimise their time usage.

As the telephone is obviously the privilege channel to inform mobile travellers not sitting at their desks, providing real-time traffic information is obviously a strong use case for speech-enabled IVRs.

Many transportation utilities are keen to provide their customers with real-time information about the availability and schedule of transportation mean they envision to use: airlines, trains, metros, bus and tramways, boats, taxis, road and speedways, …

Callers typically mention start and destination cities, or a line references. Depending on the available back-end functionality, the traffic line can suggest the caller to take another route.

Outbound notification

Outbound calls or campaigns are time consuming. Outbound campaigns follow a well-defined structure, making them ideal candidates for speech-enabled automation. Sample application areas are market research surveys, appointment setting, direct advertising, marketing campaigns or fundraising campaigns.

Practical implementations of speech-enabled outbound campaigns are:
• Order delivery confirmation
• Appointment confirmation
• Product surveys
• Event Registration

The collected information is stored for further offline handling or can be used in real time when the call is transferred to an agent. The information is used to prioritize and route the call to the best agent.

The timesaving aspect of automated speech can be used in combination with the quality of a real agent. Outbound surveys for example can be introduced by an agent, while the time consuming questionnaire is handled by the speech application.