Object lessons learned from an intelligent agents framework for telephony-based applications

Contact Center-related activities, at the business level, are common regardless of parent organizations or industries. The implementation of such activities is unlikely to change throughout the lifetime of the Contact Center and therefore, it could be the basis ("framework") upon which customized solutions can be built. Specifying such a framework is mainly a business rather than a technical problem. Programmers create "what" a system does while business analysts determine "when" the system takes certain actions. Factors influencing the framework architecture are the requirements for small, mobile, intelligent components, the ability for run-time modifications, and a reasonable foreknowledge of the technology that will minimize risks. We present a distributed, object oriented, telephony framework based on Intelligent Agents and coupled with a rules engine. The goals are reuse, platform independence, support for a distributed environment, system scalability, and adherence to standards. The framework's core is an influence engine that accepts external stimuli and based on business rules, determines the resulting actions. This engine is an intelligent agent that can be attached to any object for which intelligence is needed and thus, it is able to create an integrated, customizable solution. We discuss our architectural decisions, limitations, design choices, and experiences in development, customization, and operation of the framework. Tools and procedures used during implementation of specific solutions are also reported.