Let’s face it!
Customer contact in the enterprise is primarily all about satisfying customer needs and at minimal cost to the enterprise. In doing so, the potential for generating additional revenues immediately or, at least, retaining the customer relationship for future revenues, has always been the raison d’etre for call centers. Since the telephone has traditionally been the easiest and fastest means for consumers to initiate contact with live assistance in an enterprise organization, call center staffing is the biggest expense item to control, traditionally accounting for between 60-70% of customer contact operational costs.
One of the most important technology strategies to reduce such labor costs has been the use of automated telephone self-service applications, better known as Interactive Voice Response (IVR), using the Telephone User Interface (TUI). The TUI was based on touchtone keypad inputs and pre-recorded outputs for both speech prompts and information. IVR systems included proprietary platforms and languages for application development (“application generators”) and run-time integration with PBX/ACDs for the effective execution of these applications within the context of the enterprise telephone system and support staff.
In doing the research for a major article on migrating traditional call center technology to the next generation of IP-based “virtual” customer contact, it became very apparent that exploiting the increased flexibility of “skills-based” IP network routing will require the benefits of detailed call screening that an IVR application can provide. Knowing the identity of the caller through traditional telephony Automatic Number Identification (ANI) or Dialed Number Identification Service (DNIS) was really never good enough. We also have to know why they are calling (this time) in order to intelligently determine what available enterprise resource to access. We can’t guess about that particular criterion to determine the assignment of the caller to one of the following enterprise customer contact resources:
- A self-service business (IVR) application
- A first-level call handling agent (local ACD group, home agent, outsourced/offshore support)
- An internal expert or decision-maker (“specialist”)
- A responsible business partner
- A deferred response because there is no urgency and/or because the customer is not able to wait in queue for live assistance
So, VoIP and IP telephony need IVR capability for both intelligently screening and automated processing of customer calls, but what kind of IVR are we talking about?
Time to Migrate to IP and Speech Recognition!
The evolution of web-based online applications set the stage for telephone self-service applications to converge with those business applications, but the cost efficiencies and flexibilities of VoIP and IP telephony infrastructure are also driving the need to migrate legacy IVR technology to IP infrastructures.
On the user interface front, the limitations of the traditional TUI are being reduced by the deployment of mature speech recognition technologies for more flexible directed voice input and commands, instead of touchtone selections from voice menus.
TDM-based IVR systems have always been a proprietary combination of a development tool or “application generator” and a run-time processor that integrated with a variety of premise-based telephone switches. Although the early development languages were proprietary and required special programming expertise to deal with telephony integrations, graphical “application generator” software facilitated the creation of logical call flows and automatically generate the necessary run-time execution code.
TDM PBX/ACD integrations with IVR systems were usually replicated at every call center site, requiring maintenance and support staffing at each location. In addition, different switches required different integration implementations, as well as integration with specific business applications and databases. All of this made designing, implementing, and maintaining IVR applications difficult and expensive, especially for smaller enterprise organizations without technology support staffs.
The TUI approach for Touchtone telephones imposed a more fundamental limitation because it was only effective with structured short menus for determining caller needs. Any applications with a long list of menu choices or with many menu steps were not acceptable to callers, and they would either hang up or “zero out” to get live assistance. The recent improvement of speech recognition technology has alleviated the touchtone input problem of many or complex choices by using “directed” and flexible voice inputs.
Now that VoIP and IP telephony have changed the networking and CTI landscape, the future of telephone application self-services has to exploit the infrastructure economies of centralizing telephony applications, as well as the infrastructure commonalities of online web applications for both efficiency and consistency.
Converging Enterprise Telephone and Web Self-Services
Although visual and voice user interface logic cannot be identical in terms of prompting, menu options, error handling, and user control options, the basic flow logic should be similar and consistent. However, perhaps more importantly, at least the “back end” integrations with business applications and databases should be the same to reduce costs and maintain application consistencies.
As far as converging the user interfaces, the first step for IVR’s IP migration has been to enable the development of the voice/telephony interface within the context of web application development tools. In the last couple of years, new programming languages for speech content and telephony call control have been implemented for the standard web application development language, XML. These are VoiceXML and CCXML respectively. VoiceXML also facilitates the added use of speech recognition, which expands the potential telephone self-service applications that can be implemented with IVR technologies.
VoiceXML, initially sponsored by VoiceXML Forum founders AT&T, IBM, Lucent and Motorola, has been adopted by both new telephony technology providers like Voice Genie and Interactive Intelligence, and traditional IVR application generator providers such as Avaya, Edify, IBM, Intervoice-Brite, and Genesys, thus providing a migration path for existing applications developed with those tools. Because VoiceXML operates over the web, IVR applications can also be more easily developed and supported on a service basis.
Microsoft, Cisco, and Intel, are the heavyweights that founded the forum for a more ambitious initiative for a telephone application development language extensions, Speech Application Language Tags (SALT), based upon existing HTML, XHTM, and XML. Their approach will not only bridge the user interface modality gap between visual online applications and telephone voice applications, but will also enable multi-modal devices to exploit the combination of efficient visual output for prompting, menus, error responses, and information content, while simultaneously allowing convenient, hands-free speech input. That can also effectively capitalize upon “directed” input (through the visual prompts) and speech recognition. Obviously, this multi-modal capability will prove most practical for mobile handheld devices and will exploit the growing use of such devices by both customers and internal enterprise users for accessing self-service applications.
The supporters of VXML have recognized the need for multi-modal flexibility and are adding XHTML (XHTML+V) accommodations to their specifications.
Changing input modalities, based upon the regulatory requirements for logging certain customer contacts, can also be a practical consideration for customers using mobile multi-modal devices. Financial services providers have started to exploit instant text messaging contacts from customer desktops, rather than call recordings, in order to insure more efficient and accurate transaction records. By enabling callers to dynamically switch “transmodally” from speech to text for such sensitive data input, converged self-service applications can support selective input modalities required by a specific business function. Although speech recognition can display textual versions of voice inputs for caller confirmation, QoS issues may interfere with the accuracy of the conversion, especially with wireless devices.
Proactive “Application Buddies”
The enterprise-customer relationship has always been a two-way street. There are times when the enterprise service needs to contact a customer/user with time-sensitive notifications and alerts. This is a one-to-one communication contact between an enterprise application and an individual customer/user of an enterprise service.
Multi-modal contact flexibility will benefit from the use of SIP-based presence/modality management technologies that will dynamically determine the contact modality required by a recipient. This will be equally applicable to either human contact initiators or self-service applications that are attempting to deliver urgent, time-sensitive information to recipients. This will not be limited to one-way messages, but, once “connected,” can involve a fully interactive exchange just as if the customer/user had initiated the contact. (Think of it as outbound IVR!)
Realistically, circumstances (not preferences or rules) will dynamically determine the modality of the interaction and the proactive, self-service business application must therefore also be device and modality independent.
What Do You Think?
When do you think enterprise organizations will migrate their IVR systems to IP-IVR? Will multi-modal capability be important right away? When will enterprise customer contact operations rethink their support strategies for mobile customers? Will it happen after they have gained experience with their internal applications, i.e., “help desk,” Human Resources? With the new web-oriented IVR design tools will web application designers be able to master converged telephony versions of the same applications?
Let us know your opinions by sending them to [email protected]
IIR Conference on New VoIP-based Services for Enterprises and Consumers, April 25-27, 2005, Scottsdale, AZ
Upcoming Conference Including “Transmodal Communication”
Since writing our articles on “transmodal communication,” we have been asked to bring more perspectives that the term, “transmodal” can bring to traditional person-to-person communications.
I will be participating in an upcoming VoIP-oriented conference, primarily for carriers and service providers, which is being produced by IIR USA. Under the catchall label of “The VoIP Summit,” this conference will be discussing both consumer and business market requirements and opportunities for IP network-based communication service providers. I will be organizing a panel discussion that will discuss practical issues with new communication services for business users, such as “one-number” services.
With the growth of teleworking and wireless mobility, converged, personalized communications access is becoming a service offering that individual users in the enterprise world will be selecting and managing for themselves in, what I call, the “mass business market.” With IP interoperability, such services may now interwork in complementary ways with both wired and wireless enterprise unified communication technology to selectively support the dynamic and transmodal needs of mobile and remote accessibility.
For program information and further details, visit IIR USA’s web site at:
Art Rosenberg is a veteran of the computer and communications industry and formed The Unified-View to provide strategic consulting to technology and service providers, as well as to enterprise organizations, in migrating towards converged wired and wireless unified communications. He focuses on practical user requirements, implementation issues, and new benefits of multi-modal communication technologies for individual end users, both as a consumer and as a member of enterprise working groups. The latter includes identifying new responsibilities for enterprise communications management to support changing operational usage needs most cost-effectively.
Copyright � 2004-2005, The
Unified-View, All Rights Reserved Worldwide