# US 2014/0195378 A1 · Stage-wise analysis of text-based interactions > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2014/0195378 A1 - Document ID: US20140195378A1 - Status: Published application - Kind: A1 - Application number: 14/149,761 - Filed: 2014-01-07 - Published: 2014-07-10 - Earliest claimed: 2013-01-09 - Assignees: 24/7 Customer, Inc. - Applicants: 24/7 Customer, Inc. - Inventors: R. Mathangi SRI; Prashant V. ULLEGADDI; Vaibhav SRIVASTAVA - Web page: https://patentgrep.com/patent/US20140195378A1 ## CPC - H04M 3/5175: {Call or contact centers supervision arrangements} - G06N 20/00: Machine learning - G06Q 10/06398: {Performance of employee with respect to a job function} - G06Q 30/0601: Electronic shopping [e-shopping] - G06Q 30/0613: {using intermediate agents} - G10L 15/063: {Training} - G10L 15/142: {Hidden Markov Models [HMMs]} - H04M 3/42221: {Conversation recording systems (at the subscriber's set H04M1/656)} - H04M 3/5183: {Call or contact centers with computer-telephony arrangements} - G10L 2015/088: {Word spotting} - H04M 2203/401: Performance feedback - H04M 2203/556: Statistical analysis and interpretation - H04M: TELEPHONIC COMMUNICATION (circuits for controlling other apparatus via a telephone cable and not involving telephone switching apparatus G08) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING ## Abstract The stages of an interaction between a potential customer (the user) and a sales representative (the agent) during a sales interaction are identified to understand the interaction factors that drive sales and, by doing so, to serve the customer better and thus increase sales. Initially, a user makes contact with an agent via a communications network. During the interaction, a dropping point is reached, i.e. the point in the interaction at which either the user or the agent ends the interaction. The dropping point and other interaction factors is analyzed. Based upon such analysis, various recommendations are made to the agents to improve the user's sales experience. ## Claims ### Claim 1 A computer implemented method for user analysis, comprising: providing an interaction engine; using said interaction engine to capture an interaction between at least one agent and a user; providing an analysis engine; classifying, with said analysis engine, said interaction into one or more stages of said interaction; identifying the stage of interaction, based on the classification predicted by the analysis engine; and based on said classifying, providing recommendations to said agent to improve the user experience in the interaction. ### Claim 2 (depends on 1) The method of claim 1, further comprising: using the classification predicted by the analysis engine for offline analysis and agent performance improvement. ### Claim 3 (depends on 1) The method of claim 1, further comprising: based on said classifying, predicting any of a dropping point and recommendations leading to conversion. ### Claim 4 (depends on 1) The method of claim 1, further comprising: providing said recommendations at any of a plurality of said stages. ### Claim 5 A computer implemented method for stage-wise analysis of text-based interactions between a user and at least one agent in a sales and/or service environment, comprising: providing an interaction engine with which said user and said agent interact with each other via any of a plurality of available channels as the medium of communications; an analysis engine configured for capturing text of said interaction from said interaction engine; said analysis engine configured for classifying said interaction into a plurality of stages; based on the classification of said interaction into stages, said analysis engine configured for providing recommendations to said agent; wherein said analysis engine predicts the stage of a present interaction and modifies said interaction to achieve a favorable result. ### Claim 6 (depends on 5) The method of claim 5, wherein a mode of interaction between said agent and said user comprises any of voice-based, text-based, social network-based, and forums-based. ### Claim 7 (depends on 5) The method of claim 5, wherein said channel comprises any of a cellular based communication network, an Internet Protocol (IP) based network, a packet based communication system, a public switched telephone network (PSTN) based network, ands a voice over IP (VoIP) based network. ### Claim 8 (depends on 5) The method of claim 5, wherein said user interacts via a device that comprises any of a mobile phone, a handheld device, a tablet, a computer, and a telephone. ### Claim 9 (depends on 5) The method of claim 5, further comprising: said analysis engine configured for providing information to said agent that to push to said user in other forms of interaction and/or in one or more widgets the may be accessed by the user during said user/agent interaction. ### Claim 10 (depends on 5) The method of claim 5, further comprising: said analysis engine configured for personalizing said interaction at the user level. ### Claim 11 (depends on 5) The method of claim 5, further comprising: said analysis engine configured for optimizing flow based on any of drop-off, visitor level during a user Web journey, and interaction events. ### Claim 12 (depends on 5) The method of claim 5, further comprising: based on said recommendation, said analysis engine configured for predicting drop-off point at the start said interaction and re-evaluating the drop-off intent. ### Claim 13 (depends on 12) The method of claim 12, further comprising: said analysis engine configured for sending alerts to said agent to advise said agent whether or not sliders and/or forms should be pushed to the user. ### Claim 14 (depends on 12) The method of claim 12, further comprising: said analysis engine configured for recommending at least one channel other than a current channel as a mode of communication. ### Claim 15 (depends on 5) The method of claim 5, further comprising: said analysis engine configured for re-targeting users, once a drop-off has occurred. ### Claim 16 (depends on 5) The method of claim 5, further comprising: determining dropping point for said interaction, wherein the dropping point corresponds to a stage of the interaction at which the interaction terminates; and based on the classification and the determined dropping point, said analysis engine configured for making said recommendations. ### Claim 17 A method for analyzing interactions between a user and at least one agent, comprising: providing a control engine configured for receiving text of the interaction from said interaction engine via an interface; said control engine configured for sending a recording to a transcribing module to transcribe said recording into text upon receiving a recording of a voice-based interaction; said control engine configured for classifying each line of the interaction into one of said plurality of stages; based on the classifications of the interactions into stages, said control engine configured for determining a stage for each of a plurality of interactions; based on the classification and the determined stage, said control engine configured for providing recommendations to said agent. ### Claim 18 (depends on 17) The method of claim 17, wherein said control engine implements an algorithm which labels sequences of text into said stages. ### Claim 19 (depends on 18) The method of claim 18, wherein said control engine implements a hidden Markov model (HMM) based approach to classify said interactions into stages. ### Claim 20 (depends on 18) The method of claim 18, wherein said control engine implements a greedy sequence classifier algorithm that uses a stage predicted for a previous chat line to make a prediction for the stage of a current chat line to classify an entire conversation into stages. ### Claim 21 (depends on 17) The method of claim 17, further comprising: to distinguish between said agent and said user text, said control engine configured for learning different discriminative keywords for agent text and user text to classify said text. ### Claim 22 (depends on 17) The method of claim 17, further comprising: said control engine configured for providing recommendations for interaction flow structuring and/or restructuring. ### Claim 23 An apparatus for user analysis, comprising: an interaction engine for capturing an interaction between at least one agent and a user; and an analysis engine for classifying, with said analysis engine, said interaction into one or more stages of said interaction; said analysis engine identifying the stage of interaction, based on the classification predicted by the analysis engine; and said analysis engine providing recommendations to said agent to improve the user experience in the interaction, based on said classifying.