Systems and methods for ai-based database configuration

A method for AI-based database configuration, comprising:

• using a computing device to access a database containing historical data;
• using the computing device to access data containing one or more target values;
• deploying a trained AI program to generate a plurality of action paths based on the historical data,
• wherein the AI program comprises a neural network including at least one hidden layer between an input layer and an output layer, the at least one hidden layer configured to process the historical data through weighted sum computations at nodes;
• using the AI program to predict a probability of meeting the target value for each action path;
• selecting one of the action paths having the highest probability;
• storing the selected action path in the database;
• receiving, after a user follows the selected action path, an error signal representing a difference between the target value and an actual achieved value resulting from the user's actions along the selected action path;
• updating the historical data based on the error signal;
• retraining the AI program using the updated historical data, wherein the retraining includes adjusting weight coefficients assigned in the at least one hidden layer based on the error signal until an associated error amount is less than a predetermined threshold level to improve accuracy of probability predictions for future action paths;
• and
• generate a dashboard graphic containing a representation of the selected action path in combination with the user's progress along the selected action path.

A non-transitory computer-readable storage medium, the computer-readable storage medium including instructions that when executed by a processor, cause the processor to:

• access a database containing historical data;
• access data containing one or more target values;
• deploy a trained AI program to generate a plurality of action paths based on the historical data,
• wherein the AI program comprises a neural network including at least one hidden layer between an input layer and an output layer, the at least one hidden layer configured to process the historical data through weighted sum computations at nodes;
• use the AI program to predict a probability of meeting the target value for each action path;
• select one of the action paths having the highest probability;
• store the selected action path in the database;
• receive, after a user follows the selected action path, an error signal representing a difference between the target value and an actual achieved value resulting from the user's actions along the selected action path;
• update the historical data based on the error signal;
• retrain the AI program using the updated historical data, wherein the retraining includes adjusting weight coefficients assigned in the at least one hidden layer based on the error signal until an associated error amount is less than a predetermined threshold level to improve accuracy of probability predictions for future action paths;
• and
• generate a dashboard graphic containing a representation of the selected action path in combination with the user's progress along the selected action path.