AI-based models for assistance with near-repetitive computer-based tasks

A method, comprising:

• receiving first input from a first application programming interface (API), the first input indicating one or more input actions to a first application as monitored by the first API;
• receiving second input from a second API, the second input indicating one or more input actions to a second application as monitored by the second API, the second API being different from the first API, the second application being different from the first application;
• determining that the one or more input actions to the first and second applications have been received within a threshold amount of time;
• based on the determination, executing a prediction model to identify, from the first and second inputs, a same action being directed to different locations of a graphical user interface (GUI);
• and
• based on the identification of the same action being directed to different locations of the GUI, presenting a predicted action on a display, the predicted action being selectable by the user to provide a command to a device to autonomously perform a real action corresponding to the predicted action.

An apparatus, comprising:

• at least one computer readable storage medium (CRSM) that is not a transitory signal, the at least one CRSM comprising instructions executable by a processor system to:
• receive first input from a first application programming interface (API), the first input indicating one or more input actions to a first application as monitored by the first API;
• receive second input from a second API, the second input indicating one or more input actions to a second application as monitored by the second API, the second API being different from the first API, the second application being different from the first application;
• execute a screen scraping algorithm to track the one or more input actions to the first and second applications as a user interacts with the first and second applications;
• execute a prediction model to infer, from the one or more input actions to the first and second applications, a same type of action being performed by the user, the prediction model established at least in part by a reflective neural network;
• and
• based on the inference of the same type of action being performed by the user, present a predicted action on a display, the predicted action being selectable by the user to command the processor system to autonomously perform a real action corresponding to the predicted action.