System and method for providing automatic guidance in data flow journeys

1. A computer-implemented method of offering predictive and automated guidance to a user of software, the method comprising:

automatically initiating, via a virtual guidance system, a screen capture operation to obtain a first screenshot of activity of an experienced user of the software while the experienced user performs a first step of a first in-app task, the first screenshot including a first plurality of pixel units;

updating an electronic repository with user activity data for multiple experienced users of the software across a plurality of computing devices, the user activity data comprising screenshots, including the first screenshot, that were collected while each experienced user performed the first in-app task;

training a deep learning model to identify the first in-app task using the collected user activity data;

receiving, from a first user device, first data representing in-app behavior of a first inexperienced user of the software wherein the first data includes at least a second screenshot, and the second screenshot includes a second plurality of pixel units;

converting each of the first screenshot and the second screenshot to grayscale;

after converting each of the first screenshot and the second screenshot to grayscale, determining, using a pixel-by-pixel phase difference comparison algorithm, the first plurality of pixel units is sufficiently similar to the second plurality of pixel units so as to classify the first screenshot and the second screenshot as a match;

predicting, via the deep learning model and based on the match, that the first inexperienced user is performing the first step of the first in-app task;

automatically identifying, via the deep learning model and based on the user activity data, a sequence of in-app steps that can be used by the first inexperienced user to complete the first in-app task with greater proficiency; and

generating, via the virtual guidance system for presentation by the first user device, a step-by-step graphical model that guides the first inexperienced user through each step of the identified sequence of in-app steps thereby providing guidance about the first in-app task to the first inexperienced user.

2. The method of claim 1 , wherein the collected user activity data includes a record of where each user was located as the respective user engaged in the first in-app task.

3. The method of claim 1 , wherein the first data further includes logs and XMLs generated by the first inexperienced user while using the software.

4. The method of claim 1 , wherein the step-by-step graphical model includes an alert if a step of the sequence is associated with delays.

5. The method of claim 4 , wherein the alert includes an option to be rerouted to a different step of the sequence to avoid the step associated with delays.

6. The method of claim 1 , further comprising determining, based on the first data, that the first inexperienced user is associated with a first persona type that indicates the first inexperienced user's access level for the software, and the deep learning model bases its prediction in part on the first persona type.

7. The method of claim 1 , further comprising automatically processing the collected user activity data using feature engineering techniques including feature reduction.

8. The method of claim 1 , further comprising:

determining a second step of the sequence is currently unavailable; and

presenting a message to the first inexperienced user indicating the second step is unavailable and offering guidance about a third step that serves as an alternate to the second step in the sequence.

9. The method of claim 1 , further comprising gating the first data and using the gated first data at the deep learning model when making the prediction.

10. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to:

automatically initiate, via a virtual guidance system, a screen capture operation to obtain a first screenshot of activity of an experienced user of the software while the experienced user performs a first step of a first in-app task, the first screenshot including a first plurality of pixel units;

update an electronic repository with user activity data for multiple experienced users of the software across a plurality of computing devices, the user activity data comprising screenshots, including the first screenshot, that were collected while each experienced user performed the first in-app task;

train a deep learning model to identify the first in-app task using the collected user activity data;

receive, from a first user device, first data representing in-app behavior of a first inexperienced user of the software wherein the first data includes at least a second screenshot, and the second screenshot includes a second plurality of pixel units;

convert each of the first screenshot and the second screenshot to gray scale;

after converting each of the first screenshot and the second screenshot to gray scale, determine, using a pixel-by-pixel phase difference comparison algorithm, the first plurality of pixel units is sufficiently similar to the second plurality of pixel units so as to classify the first screenshot and the second screenshot as a match;

predict, via the deep learning model and based on the match, that the first inexperienced user is performing the first step of the first in-app task;

automatically identify, via the deep learning model and based on the user activity data, a sequence of in-app steps that can be used by the first inexperienced user to complete the first in-app task with greater proficiency; and

generate, via the virtual guidance system for presentation by the first user device, a step-by-step graphical model that guides the first inexperienced user through each step of the identified sequence of in-app steps thereby providing guidance about the first in-app task to the first inexperienced user.

11. The non-transitory computer-readable medium storing software of claim 10 , wherein the first data further includes logs and XMLs generated by the first inexperienced user while using the software.

12. The non-transitory computer-readable medium storing software of claim 10 , wherein the step-by-step graphical model includes an alert if a step of the sequence is associated with delays.

13. The non-transitory computer-readable medium storing software of claim 12 , wherein the alert includes an option to be rerouted to a different step of the sequence to avoid the step associated with delays.

14. A system for offering predictive and automated guidance to a user of software, comprising one or more computers and one or more storage devices storing instructions, the instructions when executed by the one or more computers, cause the one or more computers to:

automatically initiate, via a virtual guidance system, a screen capture operation to obtain a first screenshot of activity of an experienced user of the software while the experienced user performs a first step of a first in-app task, the first screenshot including a first plurality of pixel units;

update an electronic repository with user activity data for multiple experienced users of the software across a plurality of computing devices, the user activity data comprising screenshots, including the first screenshot, that were collected while each experienced user performed the first in-app task;

train a deep learning model to identify the first in-app task using the collected user activity data;

receive, from a first user device, first data representing in-app behavior of a first inexperienced user of the software wherein the first data includes at least a second screenshot, and the second screenshot includes a second plurality of pixel units;

convert each of the first screenshot and the second screenshot to gray scale;

after converting each of the first screenshot and the second screenshot to gray scale, determine, using a pixel-by-pixel phase difference comparison algorithm, the first plurality of pixel units is sufficiently similar to the second plurality of pixel units so as to classify the first screenshot and the second screenshot as a match;

predict, via the deep learning model and based on the match, that the first inexperienced user is performing the first step of the first in-app task;

automatically identify, via the deep learning model and based on the user activity data, a sequence of in-app steps that can be used by the first inexperienced user to complete the first in-app task with greater proficiency; and

generate, via the virtual guidance system for presentation by the first user device, a step-by-step graphical model that guides the first inexperienced user through each step of the identified sequence of in-app steps thereby providing guidance about the first in-app task to the first inexperienced user.

15. The system of claim 14 , wherein the first data further includes logs and XMLs generated by the first inexperienced user while using the software.

16. The system of claim 14 , wherein the step-by-step graphical model includes an alert if any step of the sequence is associated with delays.

17. The system of claim 14 , wherein the instructions further cause the one or more computers to determine, based on the first data, that the first inexperienced user is associated with a first persona type that indicates the first inexperienced user's access level for the software, and the first persona type is used by the deep learning model in order to filter the activity data to activity data collected for users of the same persona type.