For too many organizations, the intertwined technologies of artificial intelligence (AI) and machine learning (ML) have been long on promise but short on delivery. Increasingly, however, the fault lies not in these advanced technologies themselves, but in the challenges associated with deploying them easily and effectively to support everyday business processes.\nTrue, some AI proponents did over-promise on the field\u2019s capabilities and timetable in past decades. But a number of AI disciplines \u2013 everything from natural language processing to computer vision \u2013 have become incredibly sophisticated and powerful in recent years. AI technologies, in turn, are now powering object recognition, language translation, sentiment analysis, and a host of other use cases.\nAt the same time, ML has also matured, becoming both more capable and accessible. ML is able to monitor AI model performance over time to continually improve the model\u2019s accuracy and usefulness. In practice, organizations now often use both automated ML along with human review and updating to retrain their AI models.\nSome organizations employ data scientists to create custom AI models designed to meet specific operational needs. However, there is also a growing universe of AI models available from the open-source community as well as from individual vendors. Some offerings, such as language translation models, have already been trained on massive data sets and can be used in their generic form with little or no customization or retraining. Other models are designed to be easily customizable and updated by the organizations deploying them.\nNow, Robotic Process Automation (RPA) is emerging as a complementary technology ideally suited for transforming AI and machine learning\u2019s promise into operational reality. If well designed, an RPA platform should allow organizations to drop different AI models directly into a variety of process stages that might otherwise require human review and action.\nConsider an invoice processing workflow, which might add different AI models for extracting structured and unstructured data from invoice forms, for identifying anomalous or missing data, and for routing the invoice to the proper payment approver based on the invoice\u2019s content and payment size. By adding this type of intelligence to dozens of different workflows, organizations can greatly improve process efficiency, speed, and accuracy.\nUiPath has pioneered the marriage of RPA and AI\/machine learning with UiPath AI Center. Organizations can place AI models \u2013 built in-house or acquired from others \u2013 into AI Center, from which they can be dragged and dropped using UiPath Studio into any relevant RPA processes.\nAI Center also serves to manage AI model versioning and updates as well as for retraining models using human-validated data. In fact, any data confirmed or corrected by a human is automatically sent by robots and used to retrain the associated model.\nTo help its customers get a jump start on building intelligent RPA processes, UiPath itself has built more than 25 ML models, and also offers many intelligent models from partners via the UiPath Marketplace. UiPath Document Understanding, for example, includes models for automatically extracting data from invoices, receipts, purchase orders, and utility bills.\nBeyond the capabilities provided by AI Center and its portfolio of AI models, UiPath has embedded intelligence throughout its full UiPath Platform. The UiPath Task Mining capability, for one, leverages AI technologies to discover automation opportunities in existing organizational processes.\nFor further information about how UiPath AI Center and the library of AI models can help you further automate and improve your organization\u2019s operations, go to UiPath AI Center.