Sprint SuperMaster

Discover insights using AI to run efficient SCRUM Sprints in Jira. The app feeds contextual data to AI models to discover meaningful insights and help SCRUM Masters in making better decisions.

Comment

Inspiration

Planning and executing sprints is a tedious task, more so because any typical Jira project goes like this:

  • 100s of Jira issues, sub-issues with cross dependency and bottlenecks.
  • Several team members with different skills and skill levels.
  • Continuous and endless updates, and comments on each issue and sub-issues.
  • Jira APIs return deeply nested, complicated, and data-heavy responses.

So, unless one is willing to spend hours scanning each issue, update, comment, notification, etc., one won't be able to find bottlenecks, breakage, or delays among humongous sets of data. This generally derails sprint execution, which in turn leads to a waste of time and efforts of SCRUM Masters.

What it does

"Sprint SuperMaster" helps in discovering insights using A.I. to run efficient SCRUM Sprints in Jira.

  1. It feeds contextual data to ChatGPT to discover meaningful insights.
  2. It presents these insights as beautiful widgets in the Jira project app.
  3. It helps in sprint planning, daily SCRUM meetings, sprint reviews, and retrospectives.
  4. It provides brief action points to help SCRUM Masters make better decisions.

Insights our app helps discover

Sprint Planning

  • Step-by-step sprint planning with relevant team members, time estimates for each issue
  • Backlog issue prioritization
  • Finding who can be sprint lead based on relevant skills

Daily Standup

  • Brief summary of last updates
  • What to discuss today?
  • Questions to ask today?
  • Bottlenecks based on issue comments or discussions
  • Queries to resolve based on issue comments or discussions
  • Issues updated in the last 24 hours, or last 3 days
  • Idle issues since last 3 days, or 7 days

Sprint Review/Retrospective

  • Sprint summary
  • Key achievements
  • Key failures
  • Team performance
  • Bottlenecks faced and how to improve

How we built it

Data Analytics Approach: Jira API + AI Models: We take Jira API data and simplify it i.e. only pick contextual and useful data. This small but relevant subset of data is fed into ChatGPT and relevant queries are asked. The response is shown as widgets in the Forge project page app.

Sprint SuperMaster Forge App: It is built using the forge UI Kit and uses features like Tabs, Tables, and Fragments to build widgets. Async events are used to call ChatGPT APIs. Forge storage is used to save insights data.

Challenges we ran into

  • Tunneling issues
  • Figuring out how to run time-consuming APIs (thanks to Async Events)
  • Running multiple API queries to prepare relevant sets of data to feed to ChatGPT.
  • Widget designing using a limited set of features available in UI Kit.

Accomplishments that we're proud of

  • Fully functional Jira project app
  • Able to discover insights on the go using ChatGPT by feeding smaller contextual data.
  • App is able to help with end-to-end sprint execution.

What we learned

  • If contextual data is fed into GPT A.I., it can help unearth really meaningful insights.
  • Forge app + AI gives lots of potential to save time and effort, and eliminate the need to do redundant tasks.

What's next for Sprint SuperMaster

  • Add more widgets and insights
  • Ability to customize prompts and Jira API data schema, so that users can discover their own insights
  • Configuring additional A.I. models like Llama2.

Built With

Share this project:

Updates