Conversational access technologies originated in two distinctly different places in business enterprises. Contact centers (nee “call centers”) comprise the first site of concentrated interaction between a firm and its customers. Yet it is the corporate Web site that is having the greatest impact in cultivating the tastes, preferences and (most importantly) expectations of end users who are trying to carry out business, engage in a conversation or simply get information from the resources within an enterprise.

CAT Playing Catch-up
Contact center infrastructure has a twenty year head start on Web development. The first installation of a large-scale automated call distributor (ACD) in a customer service setting was in 1973, with the installation of a Rockwell Galaxy system at Continental Airlines. These primitive ancestors to today’s “customer interaction networks” did not exploit all the features of CAT because ACDs played the role of routing calls “to the next available agent” within the cluster of agents at their workstations.

CAT made its presence known about ten years later (in the early 1980s) as voice response units (VRUs) were installed either in front of or behind switches or automated call directors (ACDs) in enterprise call centers. Their sole purpose was to deflect some percentage of calls away from “expensive” live agents. The combination of VRUs with phone switches and computer databases were collectively referred to as “interactive voice response” (IVR) systems.

Calling IVR systems “conversational” or “interactive” at that time was too strong a statement. For one thing, caller input was confined to the digits “0” through “9” coupled with the evolving convention of using the “pound” (#) to mean “yes” or “enter” and “star” (*) to mean “go back” or in some cases “cut through to operator”. Likewise, interactivity was limited by the convention of using stored speech to prompt responses (not interactions) from callers. Worse yet, the links between enterprise information technology (IT) infrastructure and IVR systems were, for the most part feeble. The corporate mainframes that housed customer files, transaction records, product pricing, balance information and the like, saw IVR systems as a “dumb terminal.”

By emulating terminals, IVR systems relegated themselves to the role of submitting simple queries that conformed to rigid forms or screens. When the computer responded, it either told the VRU to play a stored response or responded with text strings that were rendered by robot-like text-to-speech resources (like the venerable Robbie the Robot or the nasal-laden sound of DECTalk affectionately known as “Sven the Drunken Swede”).

IP’s Role as Double Acting Catalyst
Computer-Telephone Integration (CTI) took shape during this seminal period. In the mid-1980s, as switch makers like Rockwell, IBM/Siemens and Nortel “opened” up command links that let external computers provide instructions to normally closed switches. By the early 1990’s two companies – Geotel and Genesys Labs – established themselves as the leading providers of CTI-based call management software libraries. During the subsequent years, Genesys was acquired by Alcatel and Geotel was bought by Cisco. Both evolved their product suites by adding to the library of switch interfaces and by expanding the functionality to embrace rules-based call distribution and more robust interfaces to enterprise IT. Both companies have also moved aggressively to add interactive voice response and automated speech recognition resources to their product lines.

Fast forward to 2004. VRUs and IVR systems are still the platforms of choice for call diversion from live agents and will account for over $1.5 billion in revenue in 2004. Yet the ones that merely mimic dumb terminals by responding to a limited number of instructions are heading for extinction. In their place will be commercial, off-the-shelf (COTS) platforms acting as media servers, capable of supporting conversational interactions with touchtone or spoken input and output in spoken words, streaming audio, animated graphics or video.

COTS platforms – be they Dell, HP/Compaq, IBM or Sun – can serve as IP Gateways, replace switching elements. They can also be the applications servers, database servers and Web servers that form a service-oriented architecture (SOA) that, like IP itself, easily cross the traditional boundaries between contact centers and other customer service resources.

The Internet has come on with a vengeance along two fronts. At the lower layers of the stack broadband IP is the data pipeline which is fulfilling its role as a vessel for converged communications. Yet the most profound impact is the aforementioned forming of expectation among an enterprise’s customers. The very good news, from this perspective is that, until it is rendered as speech or text on a screen in response to a customer’s query, the data in play and the application that is associated with it is indistinguishable from a Web service.

Shifting Attention to the User Experience
Users expect the same level of responsiveness, comprehensiveness and completeness of information whether they interact with a live agent, speech platform or Web site. Because considerable investment has been made over the last few years to fine tune corporate Web sites, their tastes, preferences and expectations are cultivated by their online experience. Their behaviors are slow-to-change and, in spite of the high quality of service that can be delivered over the Internet, over 70% of customer support sessions continue to be carried out over the telephone, where they are routed to either live agents or IVR systems.

However, it is clear that such systems cannot live up to caller expectations that are forged by their experience on corporate Web sites. When they confront a system with a limited number of options that accesses only a subset of the information that they have grown accustomed to reaching through the company’s Web site, they are disappointed.

Core to Achieving Business Objectives
When contact center-based resources, be they in-house agents, outsourced agents or VRUs, do not live up to caller expectation, the result is a dissatisfied customer. That’s why the largest firms in financial services, healthcare, hospitality, travel, telecommunications and technical support for electronics and communications services are making customer satisfaction as important an issue as total cost of ownership (TCO) when determining the return on investment for new contact center, CRM or call processing software and systems.

The SOA is constructed by system integrators and developers with the products and services of independent software vendors (ISVs), core speech processing “engine” providers, and the “tools” that are usually integrated development environments which increasingly include a library of reusable code or applications. The message now is that those tools must increasingly embrace Web services and the glue that brings it all together is middleware that supports Java and other standard-conformant components that operate and interoperate within IP-based networks.

Categories: Articles