Amazon Web Services (AWS) is now offering Amazon Machine Learning (AML) to developers who are looking to use predictive modeling AI in their apps and web-services. AML has the potential to save time as calculation-intense computation can consume a lot of time and server space. Amazon is already envisioning a myriad of ways developers can put Machine Learning to use. It could be used to build predictive models capable of identifying fraudulent transactions or reviews. So, review based apps like Yelp could prevent being clogged by fake reviews.
AML’s content personalization capabilities could use predictive models to recommend items like which game a user would like to play next, keeping them glued to the screen longer. And of course, the artificial intelligence could be applied to in targeted marketing, ensuring that users see are more likely to be effective.
AML could also classify documents based on their content, so it could group reviews into positive and negative categories. To prevent customers from becoming disengaged with your app or web-service, AML can identify users with a high attrition risk, and then engage them with promotions. As a nice solution to streamline customer service, AML is capable of processing feedback text from customers, and then recommending the best course of action to allay their concerns.
In Andy Jassy’s presentation for AWS at the Amazon Summit in San Francisco, he claimed that AML was able to solve a problem in 20 minutes that originally took Amazon developers 45 days to solve. Both the human developers and AML were able to improve prediction of a customer’s gender from 65% to 92% accuracy, but AML was able to complete the task in a fraction of the time.
AWS is hosting webinars to help acquaint users with its latest AML service. We can anticipate that if AML becomes widely used by developers, we’ll be seeing a lot more apps and services that are hard to turn off.