Intelligent communication routing system and method

What is this patent about?

’748 is related to the field of computer-integrated telecommunications and intelligent switching architectures. It specifically addresses the technical challenges of managing high-volume communication environments, such as call centers or data routing hubs, where multiple incoming requests must be matched with a limited pool of resources or agents. The background context highlights the limitations of traditional first-in, first-out (FIFO) or static skill-based routing, which often fail to account for the complex economic trade-offs and real-time availability fluctuations inherent in modern multi-channel communications.

The underlying idea behind ’748 is to transform communication routing from a simple rule-based selection into a dynamic aggregate utility maximization problem. Instead of just matching a caller to the first available agent with a specific skill, the system treats every potential pairing as an economic transaction with a predicted outcome. By quantifying abstract factors—such as agent proficiency, caller intent, and even the long-term value of training a trainee—into a normalized cost-benefit function, the system can perform global optimizations that prioritize high-value outcomes and operational efficiency over simple connection speed.

The claims of ’748 focus on a system and method for assigning communications by evaluating an expected economic value for each potential association between a communication and a resource. The independent claims describe a process of identifying available resources with limited capacity, calculating a score based on their availability state, and applying a communication-content dependent value function to predict the outcome of a specific pairing. This mechanism allows the processor to select an optimal routing path by comparing the estimated economic benefits across a plurality of concurrent tasks and targets.

In practice, the invention works by generating multidimensional vectors for both the communication source and the potential destination. A communication-content vector captures the specific needs and characteristics of the task, while an agent characteristic vector tracks skills, costs, and performance metrics. The system then executes a combinatorial optimization—often at a low level within the telephony server itself to minimize latency—to pair sources and targets in a way that maximizes the total utility of the entire queue, rather than just optimizing for a single call in isolation.

This approach differs from prior solutions by integrating high-level economic modeling directly into the low-level switching architecture. Traditional systems typically separate the voice switching logic from the intelligent business policy, leading to communication delays and inefficient resource use. By contrast, ’748 allows for real-time triage that accounts for opportunity costs, such as holding a highly skilled agent for a predicted high-value call rather than assigning them to a routine task. It further differentiates itself by incorporating adaptive learning, where the system updates agent profiles based on actual call outcomes to refine future routing accuracy.