The way we were: Speech technology, platforms and applications in the `Old' AT&T

Abstract

The last several years have been an exciting time at AT&T in the field of advanced speech applications for telecommunications: technical progress and platform/processor advances have enabled the identification, development and testing of a range of new services. During this period, prior to the divestiture of AT&T of Lucent Technologies and NCR, AT&T brought together, under a single corporate `roof', a research laboratory committed to advancing speech technology, business organizations building platforms to leverage this technology for telecommunications applications, and yet other business organizations with responsibility for deploying speech-enabled services to facilitate the use and reduce the cost of telecommunications services for both consumers and businesses. While this period of our corporate history has drawn to a close, we can look back to provide an overview of how technical progress, platform advances and network services needs and opportunities interacted to make speech technology an everyday experience for millions of people – and some of the lessons we learned along the way.

Résumé

Les dernières années ont été très intéressantes chez AT&T dans le domaine des applications avancées du vocal en télécommunications: des progrès techniques et des avancés au niveau des processeurs/plateformes ont permis l'identification, le développement et le test d'un ensemble de nouveaux services. Pendant cette période, et avant la séparation d'AT&T, de Lucent Technologies et NCR, AT&T rassemblait, sous un même `toit', un laboratoire de recherche engagé sur les technologies vocales avancées, des organisations d'affaires construisant des plateformes pour pousser ces technologies dans des applications en télécommunications, et d'autres organisations d'affaires ayant la responsabilité du déploiement de services à technologies vocales pour faciliter l'usage et réduire le coût des services de télécommunications pour les abonnés et les professionnels. Maintenant que cette période de notre histoire commune prend fin, nous pouvons regarder en arrière pour dresser une vue d'ensemble sur comment les progrès techniques, les avancées au niveau des plateformes et les besoins et occasions au niveau des services centralisés (à travers un réseau téléphonique) ont interagi pour que la technologie vocale devienne une expérience quotidienne pour des millions de gens – et pour décrire quelques leçons que nous avons appris au cours de ce chemin.

Introduction

Speech communication is the foundation of civilization. The use of tools built a technological society. The ability to converse with our tools has been little noted but will a have far-reaching impact. We have crossed a great technological threshold but it is anticlimactic. Science fiction has presaged it for decades and people communicate by speech so effortlessly that most do not think of hearing and understanding as a difficult problem. In the future, customers will want to interact with their machines as extensions of themselves in every possible way. Although now in an embryonic stage, the technology for humans to direct machines by speaking to them will be profoundly powerful and valuable.

This paper describes several successful services offered by AT&T based on speech recognition. A key element of developing successful services, that will not be covered, is managing the customer's expectations. When non-speech-research people hear the term automatic speech recognition their expectations are set by what they are familiar with – human speech recognition – against which the present state of the art pales in comparison. Achieving a successful service requires acquainting the buyer of the service and users of the service with the constraints of the technology so that they are not disappointed that it is not a complete human replacement.