Greetings!

Sprint 4 has now ended, the sprint went okay. There were some planed issues that the teams didn’t complete this sprint, but there were some extra issues that were added in the middle of the sprint that were completed. We didn’t reach our goals for this sprint but we got a lot of things done.

Because we switched to having a NLP model for each automation module the issues often required some changes in the NLP models and the automation modules. Thus the teams have morphed into smaller teams that focuses on one feature instead of one part of the project.

One of the things that got done this sprint is a CLI command for changing the NLP models based on language and version. This is the start of adding support for multiple languages, but the text and commands in the CLI are all still in English and we have no NLP models for other languages then English. Related to this we have also added a CLI command for installing NLP modules based on language and NLP model release, to make it easier for users to download NLP models. We have also worked a lot on implementing a system for resolving the email address only using a contacts name. Thus the system can now find the right email form just a name by searching the local contact list and the users Google Contacts. It also prompts the users with multiple alternatives if the system finds multiple similar names. The last feature that was done this sprint is a system for deleting events in the users calendar. There were some challenges with this, the main one being how to find the right event to delete using NLP. Lastly we also did some minor refactoring and minor bug fixes.

There are some features that are nearly done, but as of writhing this they are not complete. The largest and most important of those is the GUI. The design of the GUI is pretty much done but the functionality is not there. Thus implementing the functionality will have to be done the next sprint. There is also the feature of making the NLP models understand full names (last and first name), but this is far from complete.

During this sprint we also asked our product owner for a more detailed Specification Requirements and got some new requirements. The first of the new requirements is a text to speech layer that will enable the user to give voice commands to the system. The other new requirement is a chat bot that the user can interface with to get help and run the program. We have not had time figure out how we are going to achieve this using AI, but we plan to start working on it immediately.

The goals for sprint 5 will be to implement simple solutions for the two new requirements. We also plan on making it possible to update events in the users Google Calendar, so that the user don’t have to delete and create a new event. Another feature we plan to complete this next sprint is a system for piping tasks, this will enable the user to connect multiple tasks in the input. The last feature we plan to complete is the GUI.

/The RPA Tomorrow Team