1.9.0 release notes for the Engine, including the introduction of customer churn prediction template and improved data import.
At PI.EXCHANGE, we are constantly researching and exploring new ways to make AI more accessible and user-friendly to everyone. Which is why we are excited to introduce the customer churn prediction template as the major feature in this release. With this template, we hope to put the users at ease with the repetitive and laborious data science tasks, thanks to the high degree of automation and recommendation in the template. Furthermore, we have enhanced the data import process to reduce wait times for users. To provide a more consistent user experience, we have also updated the look and feel of the entire Engine.
We believe that these updates will provide our users with an improved and more streamlined experience, while also empowering them to make more informed business decisions through the AI & Analytics Engine.
Customer churn template
For release 1.9.0, the team is set out to:
-
Increase automation and recommendation to reduce time from data to prediction for the users
-
Make sure that the users can use the Engine to tackle real business problem
After some intensive development cycles, we would like to introduce the customer churn prediction template. Using the template, users can simply:
-
Submit different datasets that are not in ML-ready format
-
Specify their business problem
-
Benefit from automatic feature engineering
-
Let the Engine find and run with the most suited algorithms for their data, or select some by themselves
Once users perform these steps inside a smooth workflow, they can sit back while different processes run before coming back to review models and make predictions with the following 2 options:
-
Make a few one-off predictions to test the model’s performance on different records from the same source used for training
-
Set periodic prediction schedules to predict the likelihood of churn in the future, based on refreshed transactional data that is fetched directly from the users database, and export it to the destination of choice
Note: If you want to find out more about how use the Templated ML Solution for Customer Churn Predictions, read this article.
Improved data import
We are gradually improving the performance of data import to meet the demanding nature of the users' growing data size. In this release, we first focus on improving the speed of the process for the users to instantly see a preview of the data to check if their data is properly uploaded and parsed to quickly resolve any issue. The next improvement, which is delivered in release 1.9.1, is enabling data import to happen in the background, so that the users can freely perform other tasks in the Engine while waiting for the process to complete.
New look and feel of the Engine
For the current users, the Engine will look drastically different from the previous version. This make-over is an effort of the team to increase the consistency and, hence, usability of the Engine. Not only are the previous capabilities still available, there are more new features as mentioned above.