System and method for offering customers' appointments based on their predicted likelihood of accepting the appointment

1. A non-transitory storage medium having stored thereon instructions and deployed within a computing device, the instructions being executable by one or more processors to perform operations comprising:

responsive to an automated or autonomous trigger, training an initial predictive appointment model by the one or more processors operating in accordance with a machine learning technique based on extracted features from a training data set being a first portion of data from a dataset and testing the trained predictive appointment model based on a validation data set being a second portion of data from the dataset to generate a tested predictive appointment model;

applying the tested predictive appointment model to a dataset to generate customer appointment preference probabilities, wherein the generated customer appointment preference probabilities are maintained in a local storage;

aggregating a list of available appointments;

generating a contact list for a plurality of customers most likely to accept one of the appointments from the list of available appointments based on the customer appointment preference probabilities; and

automatically generating and transmitting electronic messages by the one or more processors over a network from the computing device to at least a computing device associated with each of the plurality of customers offering one or more available appointments and assigning an appointment of the one or more available appointments to a first customer of the plurality of customers accepting the appointment by returning a response electronic message over the network to one of the electronic messages within a prescribed time window.

2. The non-transitory storage medium of claim 1 , wherein the automated trigger includes any of (i) an addition of a new customer, (ii) an addition of a new appointment, (iii) a provision of a new appointment slot, or (iv) a modification of an existing appointment.

3. The non-transitory storage medium of claim 1 , wherein testing the trained predictive appointment model includes autonomously testing different combinations of features, weightings and coefficients until an optimum prediction accuracy is achieved to generate the tested predictive appointment model.

4. The non-transitory storage medium of claim 1 , wherein the autonomous trigger includes any one of (i) a prescribed decrease in accuracy of a current predictive appointment model or (ii) a prescribed determination of inefficiencies in the scheduling of the list of available appointments.

5. The non-transitory storage medium of claim 4 , wherein the messages provided to the plurality of customers include a selection of possible responses.

6. The non-transitory storage medium of claim 5 , wherein responsive to at least the first customer of the plurality of customers accepting the appointment, generating, by an alert generation logic, a notification message to remaining non-responsive customers of the plurality of customers that the appointment is no longer available and generating and transmitting messages by appointment scheduling logic to a second plurality of customers upon the appointment scheduling logic determining that at least one appointment of the one or more appointments is available, the second plurality of customers includes the remaining non-responsive customers assigned priority over any new customer included as part of the second plurality of customers to provide one or more alternate available appointments.

7. The non-transitory storage medium of claim 1 , wherein the electronic messages include audio signaling associated with a telephone call transmitted over a telecommunication network.

8. The non-transitory storage medium of claim 1 , wherein the electronic messages include a text message transmitted over the network.

9. The non-transitory storage medium of claim 1 , wherein the local storage corresponds to one or more databases.

10. A system for establishing a predictive appointment scheduling system, the system comprising:

one or more processors; and

a storage device having stored thereon instructions, the instructions, when executed by the one or more processors, cause the one or more processors to:

respond to an automated or autonomous trigger by at least (i) training an initial predictive appointment model in accordance with an artificial intelligence technique using extracted features from a training dataset being a first portion of data from a dataset and (ii) testing the trained predictive appointment model based on a validation dataset being a second portion of data from the dataset to generate a tested predictive appointment model;

apply the tested predictive appointment model to data accessed from a database to generate customer appointment preference probabilities, wherein the generated customer appointment preference probabilities are maintained in local storage;

generate prediction values for the customers in the system;

apply the prediction values to determine a contact list for a plurality of customers most likely to fill an available appointment based on the customer appointment preference probabilities; and

contact the plurality of customers offering the available appointment by at least automatically generating and transmitting electronic messages over a network to at least a computing device associated with each of the plurality of customers offering the available appointment and assigning the available appointment to a first customer of the plurality of customers accepting the appointment by returning a response electronic message over the network to one of the electronic messages within a prescribed time window.

11. The system of claim 10 , wherein the electronic messages include a text message transmitted over the network.

12. The system of claim 11 , wherein the artificial intelligence technique includes a machine learning technique, the artificial intelligence technique or the machine learning technique includes Random Forest, Logistic Regression, Naive Bayes (Multinomial), k-Nearest Neighbors, Deep Neural Networks, “Beta-geometric/NBD” (BG/NBD), and Support Vector Machines.

13. The system of claim 10 , wherein the data in the database includes at least one of customer information, customer scheduling data, predicted customer preferences, or previously tested models.

14. The system of claim 10 , wherein responsive to one of the plurality of customers accepting the appointment, generating and transmitting electronic messages to a second plurality of customers different than the plurality of customers of one or more alternative available appointments, the second plurality of customers assigning a priority to a remainder of the plurality of customers being part of the second plurality of customers.

15. The system of claim 10 , wherein responsive to a customer of the plurality of customers accepting the available appointment after expiration of the prescribed time window after the available appointment has been assigned to the first customer, the appointment scheduling logic generating one or more alternate available appointments for selection by the customer, the one or more alternative available appointments being selected based on the predictive values associated with the customer.

16. The system of claim 10 , wherein generating prediction values includes generating prediction values for at least one of day, time and period.

17. The system of claim 10 , wherein the electronic messages include audio signaling associated with a telephone call transmitted over a telecommunication network.

18. The system of claim 10 , wherein the tested predictive appointment model includes weightings and coefficients associated with the extracted features, the weightings and coefficients are developed during the testing of the trained predictive appointment model.

19. A method for establishing a predictive appointment scheduler system, the method comprising:

responsive to an automated or autonomous trigger, training an initial predictive appointment model operating in accordance with a machine learning technique based on extracted features from a training dataset and testing the trained predictive appointment model based on a validation dataset to generate a tested predictive appointment model;

applying the tested predictive appointment model to a dataset to generate customer appointment preference probabilities, wherein the generated customer appointment preference probabilities are maintained in a local storage;

aggregating a list of available appointments;

generating a contact list for a plurality of customers most likely to accept one of the appointments from the list of available appointments based on the customer appointment preference probabilities; and

automatically generating and transmitting electronic messages over a network to at least a computing device associated with each of the plurality of customers offering one or more available appointments and assigning an appointment of the one or more available appointments to a first customer of the plurality of customers accepting the appointment by returning a response electronic message over the network to one of the electronic messages within a prescribed time window.

20. The method of claim 19 , wherein the automated trigger includes any of (i) an addition of a new customer, (ii) an addition of a new appointment, (iii) a provision of a new appointment slot, or (iv) a modification of an existing appointment.

21. The method of claim 19 , wherein a proportion of data associated with the training dataset to data associated with the validation dataset is a ratio of 80%:20%.

22. The method of claim 19 , wherein the trained predictive model is generated using the machine learning technique or any autonomous technique, including at least one of Random Forest, Logistic Regression, Naive Bayes (Multinomial), k-Nearest Neighbors, Deep Neural Networks, “Beta-geometric/NBD” (BG/NBD), and Support Vector Machines techniques.

23. The method of claim 19 , wherein the autonomous trigger includes any one of (i) a prescribed decrease in accuracy of a current predictive appointment model in use or (ii) a prescribed determination of inefficiencies in the scheduling of the list of available appointments.

24. The method of claim 19 , wherein responsive to a customer of the plurality of customers accepting the appointment after expiration of the prescribed time window after the appointment has been assigned to the first customer, the method further comprising:

generating, by appointment scheduling logic, one or more alternate available appointments for selection by the customer, the one or more alternative available appointments being selected customer appointment preference probabilities associated with the customer.

25. The method of claim 19 , wherein the generating and transmitting of the electronic messages over the network comprises generating and transmitting audio signaling associated with a telephone call transmitted over a telecommunication network.

26. The method of claim 19 , wherein the generating and transmitting of the electronic messages over the network comprises generating and transmitting a text message transmitted over the network.

27. The method of claim 19 , wherein the local storage corresponds to one or more databases.

28. The method of claim 19 , wherein the tested predictive appointment model includes weightings and coefficients associated with the extracted features, the weightings and coefficients are developed during the testing of the trained predictive appointment model.