AI System for Agent Performance Characterization

ChangeBridge: Patent Apps - Business Methods (G06Q)

Detected March 26th, 2026

Summary

The USPTO has published a patent application (US20260087439A1) for an AI system designed to characterize agent performance using communication transcripts. The application, filed by Verizon Patent and Licensing Inc., details a method for analyzing historical agent-customer interactions to match new communications to suitable agents.

View original document View source feed page

What changed

This document is a patent application (US20260087439A1) filed with the USPTO by Verizon Patent and Licensing Inc. It describes an AI-enabled system for characterizing agent performance by analyzing communication transcripts. The system generates historical agent features and uses current dynamics to match new communications to appropriate agents, aiming to improve efficiency and customer service.

As this is a patent application, it does not impose direct regulatory obligations on companies. However, it signals potential future technological advancements and intellectual property claims in the area of AI-driven customer service and performance management. Companies operating in this space, particularly those utilizing AI for agent management or customer interaction routing, should be aware of such patent filings as they may impact future product development or licensing strategies.

Source document (simplified)

← USPTO Patent Applications

METHOD AND SYSTEM FOR AI-ENABLED CHARACTERIZATION OF PERFORMANCE AND APPLICATIONS THEREOF

Application US20260087439A1 Kind: A1 Mar 26, 2026

Assignee

Verizon Patent and Licensing Inc.

Inventors

Nilanjan Das Khan, Tushar Singh, Goutham Kumar Chikoti, R Vishnu

Abstract

The present teaching relates to AI-enabled characterization of agent performance. Transcripts of historic communications between agents and customers are analyzed by AI-enabled models to generate historic agent features to characterize agents'historic performance. When a new communication associated with an intent is initiated, candidate agents are identified according to the intent and their historic agent features. Current dynamics of each candidate agent are determined via AI-enabled models. A matching agent is selected based on the historic agent features and current dynamics of the candidate agents so that the new communication is directed to the matching agent.