RPA Network Request Generation to Bypass Slow Web UI Tasks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robotic process automation (RPA) tasks interacting with web pages face inefficiencies due to long load times for executing UI commands, as they often require loading UI elements for data input, which can slow down processing.
Innovation Solution
A computer-implemented method using machine learning, specifically recurrent neural networks (RNNs) and Long Short Term Memory (LSTM) or Gated Recurring Unit (GRU) neural networks, generates network requests from runtime variables of RPA bots executing client-side web application code, allowing these bots to bypass execution of RPA commands and directly send network requests with payload data to web applications for processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If RPA bots execute UI commands to interact with web applications, then they can perform automated tasks through user interface interactions, but processing speed is reduced due to long load times for loading UI elements
Solution Approach 1:
The patent extracts the essential function of UI interaction (sending data to the web application) from the UI command execution process. Instead of loading UI elements and simulating user interactions, the system directly generates and sends network requests containing the necessary payload data, eliminating the slow UI loading step while preserving the core automation capability
Solution Approach 2:
The patent replaces the mechanical UI interaction system (loading UI elements, simulating keyboard/mouse actions) with a direct network communication system. The machine learning model generates network requests that bypass the UI layer entirely, substituting the mechanical interaction process with a more efficient electronic communication process
2Ease of operation
If RPA bots load UI elements for data input, then they can interact with web application controls, but processing time increases due to sequential loading requirements
Solution Approach 1:
The system performs preliminary action by using the machine learning model to predict and generate the complete network request payload before any UI elements are loaded. The runtime variables and payload data are prepared in advance based on the bot's intended action, eliminating the need to wait for UI elements to load before determining what data to send
Solution Approach 2:
The patent extracts the payload data generation process from the UI element loading process. Instead of waiting for UI elements to load and then extracting data from them, the system directly generates the necessary payload data through the machine learning model, separating data preparation from UI rendering
3Extent of automation
If RPA commands are executed to send data to web applications, then tasks can be automated through standard protocols, but execution speed is reduced due to command interpretation overhead
Solution Approach 1:
The patent creates a copy of the essential communication function by generating network requests that replicate the data transmission purpose of UI commands. Instead of executing the full UI command sequence, the system creates and sends a simplified network request that copies only the necessary data transmission functionality, eliminating interpretation overhead
Solution Approach 2:
The machine learning model acts as an intermediary between the bot's intent and the web application. It receives runtime variables from the bot, processes them to determine the appropriate action, and generates the corresponding network request, mediating the communication without requiring standard UI command execution
Data Source
AI summary
Aspects of the present disclosure relate generally to robotic process automation (RPA) and, more particularly, to systems, computer program products, and methods of accelerating RPA tasks that interact with web pages with machine learning generation of network requests. For example, a computer-implemented method includes receiving, by a processor, plural runtime variables of a bot executing client-side web application code; generating, by the processor, a network request from a machine learning model given input of the plural runtime variables of the bot; and sending, by the processor, the network request generated from the machine learning model to the bot.