This month we went back to school and had the chance to use AI as our assistant!

With a simplified "ACME School" organization, where members are divided in 2 subgroups: teachers and students, we worked on getting ready for the school period closure. As the teacher we had to give feedback to students based on their grades.

Our "PeriodFeedback" process allowed us to discover how to create a task with parallel multi-instantiation. Indeed, we needed to launch at the same time the feedback for all students in parallel. We also attributed automatically the tasks to the right person, i.e. the teacher in charge of those students. We did this with a Single User actor filter (available on the Bonita Marketplace) on the "FeedbackSubmission" process.

In this process as well we used Artificial Intelligence as an assistant for writing feedback based on our raw notes. The idea is that the teacher can give all the input as notes and the assistant will help him/her write a more elaborated text base on these notes for the student. For achieving this, we created a custom connector from scratch to connect to Open AI. We published it and added it to our project as a Maven dependency.

Useful links shown during this session

• Our Bonita project on Github: https://github.com/Bonitasoft-Community/school-feedback-example
• Our custom connector code on Github: https://github.com/thaliacruz-bonita/custom-connector-openai-example
• Bonita Documentation on connectors: https://documentation.bonitasoft.com/bonita/2022.2/runtime/connectors-execution#_what_are_connectors
• Maven getting started: https://maven.apache.org/guides/getting-started/maven-in-five-minutes.html
• Retrofit: https://square.github.io/retrofit/
• OpenAI Completions API Reference: https://platform.openai.com/docs/api-reference/completions/create

Did you miss it?

Don't worry!

WATCH THE REPLAY IN ENGLISH

VE LA GRABACIÓN EN ESPAÑOL

REGARDEZ LE REPLAY EN FRANÇAIS