Technology and Data News and Events

Enterprise Resource Planning Finance System

A closer look at Oracle Adaptive Intelligent Apps

by Mike Jones in Artificial Intelligence, Machine Learning, Oracle 30/05/2019

In a recent report, Melissa Boxer, the VP of Oracle explains how Oracle Adaptive Intelligent Applications can remove the repetitive and tedious ERP and supply chain jobs from people’s work.

Enterprise Resource Planning Finance System
Over the last few years, Oracle has increased its investment into implementing AI and Machine Learning capabilities within a range of their applications, in particular, the Oracle Adaptive Intelligent Applications. Boxer explains that its Adaptive Intelligent Apps are focused on creating next-generation smart innovative applications based on big data generated within the Oracle Cloud. Boxer highlights that AI and ML can be integrated into Oracle Cloud-based applications in various methods. She explains how the automation of tedious and repetitive jobs such as invoice matching or approving expenses are tasks that can be handled by AI/ML tools within the cloud today.

Boxer explains that ML can manage specific users cases within conventional business activities, such as procurement and accounts payable. Oracle then expand on this potential by using Oracle’s ML algorithms to enhance the entire process.

For Oracle, ML requires accurate and relevant data to generate the best results. It’s quite clear that relying on poor data for ML models will mean poor quality results. Data fed to the Oracle Adaptive Intelligent Apps utilises a blend of first and third party data pools. Oracle uses tuning processes, using training data and ML algorithms ensuring customers receive the benefits of instant value when the application is activated. The data records are consistently added to the Oracle ML learning systems, enabling models to be constantly refined based on customer data. The concept behind these models is to reduce the reliance on having a pool of data scientists, which for many businesses is simply not economically viable.

Boxer points out that there is no assumption that the AI developed for standard projects will work for all. The system needs to gain an understanding of the customer business, which involves learning from customer actions and their response to recommendations. Through each recommendation, the system learns from the customer response. The information generated supports ML algorithms and ensuring they are continuously improved.

For Oracle, AI is integrated into applications that existing customers use, meaning they don’t have to become familiar with another platform. Chabot engagement is a growing system that is developed to provide an enhanced mobile experience and enable people to perform a task by responding to several simple questions. In terms of expense reporting, this applied process can save thousands of hours for both the employee and managers, who presumably have better things to be doing than collating expense reports.