Systems and methods for facilitating dialogue mining

• US 9,661,067 B2
• Filed: 12/23/2014
• Issued: 05/23/2017
• Est. Priority Date: 12/23/2013
• Status: Active Grant

• Current Assignee: [24]7.ai, Inc.
• Original Assignee: 24/7 Customer Incorporated
• Inventors: Kannan, Pallipuram V., Seth, Suchana, Singh, Bhupinder, Ramachandran, Mathangi Sri, Vijayaraghavan, Ravi
• Primary Examiner(s): Jacob, Ajith
• Application Number:
  US14/580,859
• Publication Number:
  US 20150178371A1
• Time in Patent Office: 882 Days
• CPC Class Codes: G06F 16/2428 Query predicate definition using graphical user interfaces, including menus and forms G06F16/2423 takes precedence
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 40/117 Tagging; Marking up details of markup languages G06F40/143; Designating a block; Setting of attributes style sheets, e.g. eXtensible Stylesheet Language Transformation [XSLT], G06F40/154
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]
• CPC Class Codes: G06Q 30/016 After-sales
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 51/04 Real-time or near real-time messaging, e.g. instant messaging [IM]
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

The disclosure is related to mining of text to derive information from the text that is useful for a variety of purposes. The text mining process can be implemented in a service oriented industry such as a call center, where a customer and an agent engage in a dialog, e.g., to discuss product/service related issues. The messages in dialogues between the customers and the agents are tagged with features that describe an aspect of the conversation. The text mining process can mine various dialogues and identify a set of features and messages based on prediction algorithms. The identified set of features and messages can be used to infer an intent of a particular customer for contacting the agent, and to generate a recommendation based on the determined intent.

Claims

1. A method for facilitating a communication system to generate a set of messages for an agent to respond to a customer, comprising:

converting a plurality of messages exchanged between the customer and the agent during a dialog to a plurality of structured messages, the plurality of messages exchanged using a chat application that is configured to execute on a first computing device associated with the agent and a second computing device associated with the customer, the converting comprising;

tagging the plurality of messages, at least by the agent, with one or more features of a plurality of features, wherein each of the features depict one or more extracted aspects of the dialog; and

storing the plurality of messages with their tagged features as the plurality of structured messages at a storage system associated with the communication system, the storage system comprising a second plurality of structured messages that represent a second plurality of messages exchanged between a plurality of agents and a plurality of customers, the second plurality of structured messages tagged with a second plurality of features;

generating, by the communication system, a first message depicting an intent of the customer for contacting the agent, the generating comprising;

extracting, based on one or more prediction techniques, a group of features and a group of messages from the plurality of structured messages, the second plurality structured messages, the plurality of features and the second plurality of features using a text mining process to generate mined data;

determining the intent of the customer as a function of the mined data and a user profile of the customer; and

generating, by the communication system, based on the determined intent, a second message containing one or more recommendations for responding to the customer.

2. The method of claim 1, said tagging further comprising:

automatically tagging using any of information extraction algorithms, named entity recognition for person name and location name, spell checking algorithms, string similarity, emotion features, persona features, nps drivers, dialog stage, canned phrases, check marks, radio buttons, voice referral, chat transfer, and chat disconnects.

3. The method of claim 1, further comprising:

exposing different features to different target groups for tagging.

4. The method of claim 2, said tagging further comprising:

collaborative tagging for validating and/or correcting the automatic tagging.

5. The method of claim 1, further comprising:

integrating incremental learning or active learning models to showcase a reduced set of categories for selection instead of an entire list.

6. The method of claim 5, wherein said reduced set comprises a most likely set of categories as predicted by a model.

7. The method of claim 5, wherein categories with least confidence scores from the model are showcased for selection to tag more data to improve model performance for those categories over time.

8. The method of claim 4, said collaborative tagging further comprising:

agents, customers, supervisors, and QA teams using any of a plurality of widgets for collaborative tagging to improve the speed, efficiency, and accuracy of data collection.

9. The method of claim 1, further comprising:

event based tagging to capture features as structured data, wherein said events comprise any of a number of deletions, backspaces used by the agent, the number of copy paste events that occurred for an agent recommendation, and feedback button clicks on agent recommendations.

10. The method of claim 1, further comprising:

tracking cross-sells and outcomes with text categorization models that predict a cross-sell and an outcome type in a chat.

11. The method of claim 1, further comprising:

tracking cross-sells and outcomes with a collaborative platform.