Intelligent call analysis and orchestration using machine learning, generative artificial intelligence (AI), and distributed ledgers

1 . A computing platform comprising:

at least one processor;

a communication interface communicatively coupled to the at least one processor; and

memory storing computer-readable instructions that, when executed by the at least one processor, cause the computing platform to:

train, using standard operating procedures (SOP) and historical agent response information, a first machine learning model comprising a first neural network, wherein training the first machine learning model configures the first machine learning model to output, based on live interaction information, a solution deviation score indicating a degree to which agents are complying with a corresponding SOP;

train, using voice spectrograms for historical agent interactions, a second machine learning model comprising a second neural network, wherein training the second machine learning model configures the second machine learning model to output, based on the live interaction information, an emotional deviation score indicating an emotional state of a corresponding agent;

detect a live interaction between an agent and a customer;

input information of the live interaction into the first machine learning model and the second machine learning model to produce, via execution of the first neural network and the second neural network, a first solution deviation score and a first emotional deviation score;

compare the first solution deviation score to a solution deviation threshold and the first emotional deviation score to an emotional deviation threshold;

based on detecting that one or more of the solution deviation threshold or the emotional deviation threshold are met or exceeded, initiate one or more corrective actions, wherein:

the one or more corrective actions comprise an agent based action and a system based action,

the agent based action comprises providing live feedback to the agent to facilitate the live interaction via a homomorphic encryption layer, wherein use of the homomorphic encryption layer ensures privacy of the live feedback,

the system based action comprises updating a first SOP corresponding to the live interaction and updating one or more interface elements corresponding to the live interaction, wherein updating the one or more interface elements comprises moving information to a new tab or subpage of an interface comprising the one or more interface elements, and

updating the first SOP comprises:

generating, based on the information of the live interaction, a prompt for a generative artificial intelligence (AI) model, and

inputting, into the generative AI model, the prompt, wherein inputting the prompt into the generative AI model causes the generative AI model to produce an updated SOP, wherein the first SOP is replaced by the updated SOP;

update, via a dynamic feedback loop and based on the first solution deviation score and the first emotional deviation score, the first machine learning model and the second machine learning model, wherein the update improves accuracy of the first machine learning model and the second machine learning model in detecting performance deviations, and wherein updating the first machine learning model and the second machine learning model is in response to detecting that the accuracy is less than an accuracy threshold;

compare, after updating the first machine learning model and the second machine learning model, a subsequent accuracy of the first machine learning model and the second machine learning model to the accuracy threshold; and

based on detecting that the subsequent accuracy meets or exceeds the accuracy threshold, pause updates to the first machine learning model and the second machine learning model.

2 . The computing platform of claim 1 , wherein the first machine learning model is further trained using a historical voice event interaction sequence log, wherein historical voice interaction metadata is mapped to historical telemetry information to produce the historical voice event interaction sequence log.

3 . The computing platform of claim 1 , wherein initiating the one or more corrective actions comprises deriving, based on one or more of the first solution deviation score or the first emotional deviation score, the one or more corrective actions, wherein the one or more corrective actions are initiated to address a particular issue reflected by the first solution deviation score or the first emotional deviation score.

4 . The computing platform of claim 1 , wherein the agent based action comprises providing live feedback to the agent to facilitate the live interaction.

5 . The computing platform of claim 1 , wherein the generative AI model comprises one or more of: a knowledge graph, a long short term memory (LSTM) model, a transformer model, or a generative adversarial networks (GAN) model.

6 . The computing platform of claim 1 , wherein the memory stores additional computer-readable instructions that, when executed by the at least one processor, cause the computing platform to:

identify whether or not a number of agents experiencing a live interaction problem exceeds a predetermined agent threshold, wherein updating the first SOP or the one or more interface elements corresponding to the live interaction comprises updating, in response to identifying that the predetermined agent threshold is exceeded, the first SOP or the one or more interface elements, wherein the one or more corrective actions comprise the system based action.

7 . A method comprising:

at a computing platform comprising at least one processor, a communication interface, and memory:

training, using standard operating procedures (SOP) and historical agent response information, a first machine learning model comprising a first neural network, wherein training the first machine learning model configures the first machine learning model to output, based on live interaction information, a solution deviation score indicating a degree to which agents are complying with a corresponding SOP;

training, using voice spectrograms for historical agent interactions, a second machine learning model comprising a second neural network, wherein training the second machine learning model configures the second machine learning model to output, based on the live interaction information, an emotional deviation score indicating an emotional state of a corresponding agent;

detecting a live interaction between an agent and a customer;

inputting information of the live interaction into the first machine learning model and the second machine learning model to produce, via execution of the first neural network and the second neural network, a first solution deviation score and a first emotional deviation score;

comparing the first solution deviation score to a solution deviation threshold and the first emotional deviation score to an emotional deviation threshold;

based on detecting that one or more of the solution deviation threshold or the emotional deviation threshold are met or exceeded, initiating one or more corrective actions, wherein:

the one or more corrective actions comprise an agent based action or and a system based action,

the agent based action comprises providing live feedback to the agent to facilitate the live interaction via a homomorphic encryption layer, wherein use of the homomorphic encryption layer ensures privacy of the live feedback,

the system based action comprises: updating a first SOP corresponding to the live interaction, ander updating one or more interface elements corresponding to the live interaction, wherein updating the one or more interface elements comprises moving information to a new tab or subpage of an interface comprising the one or more interface elements, and

updating the first SOP comprises:

generating, based on the information of the live interaction, a prompt for a generative artificial intelligence (AI) model, and

inputting, into the generative AI model, the prompt, wherein inputting the prompt into the generative AI model causes the generative AI model to produce an updated SOP, wherein the first SOP is replaced by the updated SOP;

updating, via a dynamic feedback loop and based on the first solution deviation score and the first emotional deviation score, the first machine learning model and the second machine learning model, wherein the update improves accuracy of the first machine learning model and the second machine learning model in detecting performance deviations, and wherein updating the first machine learning model and the second machine learning model is in response to detecting that the accuracy is less than an accuracy threshold;

comparing, after updating the first machine learning model and the second machine learning model, a subsequent accuracy of the first machine learning model and the second machine learning model to the accuracy threshold; and

based on detecting that the subsequent accuracy meets or exceeds the accuracy threshold, pausing updates to the first machine learning model and the second machine learning model.

8 . The method of claim 7 , wherein the first machine learning model is further trained using a historical voice event interaction sequence log, wherein historical voice interaction metadata is mapped to historical telemetry information to produce the historical voice event interaction sequence log.

9 . The method of claim 7 , wherein initiating the one or more corrective actions comprises deriving, based on one or more of the first solution deviation score or the first emotional deviation score, the one or more corrective actions, wherein the one or more corrective actions are initiated to address a particular issue reflected by the first solution deviation score or the first emotional deviation score.

10 . The method of claim 7 , wherein the agent based action comprises providing live feedback to the agent to facilitate the live interaction.

11 . The method of claim 7 , wherein the generative AI model comprises one or more of: a knowledge graph, a long short term memory (LSTM) model, a transformer model, or a generative adversarial networks (GAN) model.

12 . One or more non-transitory computer-readable media storing instructions that, when executed by a computing platform comprising at least one processor, a communication interface, and memory, cause the computing platform to:

train, using standard operating procedures (SOP) and historical agent response information, a first machine learning model comprising a first neural network, wherein training the first machine learning model configures the first machine learning model to output, based on live interaction information, a solution deviation score indicating a degree to which agents are complying with a corresponding SOP;

train, using voice spectrograms for historical agent interactions, a second machine learning model comprising a second neural network, wherein training the second machine learning model configures the second machine learning model to output, based on the live interaction information, an emotional deviation score indicating an emotional state of a corresponding agent;

detect a live interaction between an agent and a customer;

input information of the live interaction into the first machine learning model and the second machine learning model to produce, via execution of the first neural network and the second neural network, a first solution deviation score and a first emotional deviation score;

compare the first solution deviation score to a solution deviation threshold and the first emotional deviation score to an emotional deviation threshold;

based on detecting that one or more of the solution deviation threshold or the emotional deviation threshold are met or exceeded, initiate one or more corrective actions, wherein:

the one or more corrective actions comprise an agent based action and a system based action,

the agent based action comprises providing live feedback to the agent to facilitate the live interaction via a homomorphic encryption layer, wherein use of the homomorphic encryption layer ensures privacy of the live feedback,

the system based action comprises: updating a first SOP corresponding to the live interaction, and updating one or more interface elements corresponding to the live interaction, wherein updating the one or more interface elements comprises moving information to a new tab or subpage of an interface comprising the one or more interface elements, and

updating the first SOP comprises:

generating, based on the information of the live interaction, a prompt for a generative artificial intelligence (AI) model, and

inputting, into the generative AI model, the prompt, wherein inputting the prompt into the generative AI model causes the generative AI model to produce an updated SOP, wherein the first SOP is replaced by the updated SOP;

update, via a dynamic feedback loop and based on the first solution deviation score and the first emotional deviation score, the first machine learning model and the second machine learning model, wherein the update improves accuracy of the first machine learning model and the second machine learning model in detecting performance deviations, and wherein updating the first machine learning model and the second machine learning model is in response to detecting that the accuracy is less than an accuracy threshold;

compare, after updating the first machine learning model and the second machine learning model, a subsequent accuracy of the first machine learning model and the second machine learning model to the accuracy threshold; and

based on detecting that the subsequent accuracy meets or exceeds the accuracy threshold, pause updates to the first machine learning model and the second machine learning model.

13 . The non-transitory computer-readable media of claim 12 , wherein the first machine learning model is further trained using a historical voice event interaction sequence log, wherein historical voice interaction metadata is mapped to historical telemetry information to produce the historical voice event interaction sequence log.

14 . The non-transitory computer-readable media of claim 12 , wherein initiating the one or more corrective actions comprises deriving, based on one or more of the first solution deviation score or the first emotional deviation score, the one or more corrective actions, wherein the one or more corrective actions are initiated to address a particular issue reflected by the first solution deviation score or the first emotional deviation score.

15 . The non-transitory computer-readable media of claim 12 , wherein the agent based action comprises providing live feedback to the agent to facilitate the live interaction.