Standup Assistant

Standup Assistant generated daily summary for yesterday
Standup Assistant generated more sumaries of the team updates

Inspiration

Our everyday standups go as in every developer's team. Each teammate tries to remember what was done yesterday. We wanted to streamline daily standup meetings by automating the process of generating insights from Jira updates. This not only saves time but also ensures team-wide awareness of project progress.

What it does

"Standup Assistant" is a Jira app that takes issue updates and converts them into concise daily standup speeches. It gathers data from Jira, analyzes the updates, comments, and activities, and generates a structured summary of the team's progress.

How we built it

We built "Standup Assistant" using Atlassian Forge as a frontend and main part. For the backend, we do have AWS lambda with Python and OpenAI chatbot. Here's a high-level overview of how we built the project:

Data Collection: We used the Jira API to fetch issue updates, comments, and other relevant data.
Data Processing: Prepare big JSON files and structure to be clear for the AI. We wrote custom algorithms to categorize Jira updates into "yesterday," "today," and "obstacles" sections.
Request to AI: We designed a middle-layer on Python which adds Prompt and additional information for the request to the AI and process output after. Unfortunatelly, AI is not yet super reliable, so it requires result review, and multiple tries are still needed to have a good result.
User Interface: The app has an intuitive interface for team members to present structured and summarized results on the Project page.
Integration with Jira: We ensured a seamless and reliable integration with Jira even for long-running requests.

Challenges we ran into

Data Parsing: Parsing and categorizing Jira updates was challenging due to varied formats.
Integration: Ensuring smooth integration with Jira, Forge limitations, and data synchronization, was meticulous work.
Natural Language Processing: It was not the easiest task to get the needed result from AI. As it processes each request in a unique manner.

Accomplishments that we're proud of

We are proud of several accomplishments:

It's always cool to make a big thing TOGETHER as a team.
Olga learned the possibilities and limitations of the Forge.
Adrian had a wonderful time talking with the AI.
Roman just had some fun

What we learned

Building "Standup Assistant" taught us:

Possibilities and limitations of the Forge.
Challenges of natural language processing.
The possibilities of integrating AI into real life.

What's next for Standup Assistant

In the future, we plan to:

Add customization options for different standup formats, like Scrum or Kanban.
Make more apps more stable.
Teach AI to better understand the Jira project specifically to create better-structured responses.
Give end-users the possibility to customize the output, to align the summary style with internal communication style

Done by Wombats Corp