UiPath is one of the leading RPA platforms and has been recognized as a Magic Quadrant leader in RPA by Gartner in 2019 and 2020. To maintain its stronghold as the industry leader, UiPath launched AI Fabric, a ground breaking service aimed at incorporating machine learning models within RPA workflows.
RPA drives automation by automating repetitive tasks like invoice processing, task management, application integration, data integration, and so on. These actions are straightforward and don't require critical thinking. Adding AI capabilities into RPA workflows opens up new automation potential like making informed decisions, predictive intelligence, and so on. Generally, AI in association with RPA can address many real-world problems.
AI Fabric deploys machine learning capabilities within RPA workflows. Developers can choose the requisite AI and ML Model and just drag and drop into the RPA Workflows via UiPath. AI Fabric use cases are similar to RPA uses cases that add an additional layer of intelligent business automation.
The efficiency of RPA workflows to deliver results depends on the accuracy of the input data. AI Fabric simplifies the process by providing necessary information to develop and evolve the AI model according to varying business requirements and context. Thereby ensuring that the developed models are smart enough to handle uncertainties, exceptions, and other challenges. To enhance the efficiency of the RPA workflows, AI Fabric brings together the developer, analyst, business professional, and data scientist.
The Machine Learning models in AI Fabric consist of 3 parts: ML Packages, ML Skills, and ML Logs.
ML Packages encompass the Machine Learning model code and are available within the RPA workflows in Orchestrator and UiPath Studio. These packages can be managed in Orchestrator through ML Packages and allows us to view packages in available versions, statuses, and changelogs. We can upload new package versions and also delete unused or undeployed packages.
ML Skills are ready-to-use in RPA workflows. Upon deploying ML packages or OS packages as ML skills, these can be directly used in RPA workflows. The Process Controller, Data Scientist, or RPA Developer are authorized user roles to handle these deployments. The ML Skills page provides detailed information (like status, package name, version, and so on) about the models deployed.
AI Fabric deploys skills created within the machine language skill. After deployment, optimizing security checks, installing various dependencies, setting up a network within the tenant namespaces, and finally checking the overall performance of the Machine Language Skill. Machine Language Logs are consolidated to Machine Language related events, such as package validation, deployment, and prediction errors. Each event is added to start and finish time logs and checked for errors. Logs can be managed via the ML Logs page within Orchestrator.