JiraTales

JiraTales - Waiting for the first closed issue
JiraTales - As tasks are done, the collab tale unfolds
Jira Tales - Your team races against time to shape the outcome.

Inspiration

While trying to come up with ideas for this great competition, we were reminded of a phenomenal genre that we have been dreaming of realizing a variation of in the digital world for years: Our childhood's (or let's say 80s' and 90s') "choose your path" books. (i) Today, "storytellers" are followed with greater interest than others, both in business and in social life. (ii) At work, people often need a source of excitement and motivation for tasks that seem simple (but can be completed inefficiently due to procrastination, etc.). (iii) We just thought that by telling them interesting "tales" about their current work (and giving them the chance to decide the fate of the tale), we can make "resolving issues" in Jira much more fun and efficient. And so JiraTales was born.

What it does

JiraTales' AI appears to Jira users in the guise of a wise, old wizard and begins to tell them an intriguing "tale": An urban mystery in NY, an epic narrative set in medieval times, or any other thing that arouses curiosity. As users resolve project issues in Jira, JiraTales enriches the tale in parallel. The app gives the user who resolved the Jira issue the chance to determine the path the tale will take (she can leave its fate to JiraTales if she wishes). When all the issues of a project are closed and the project is successfully completed, it also means that team members have collaboratively brought a great tale to an end.

How we built it

The app's tale generating dynamics became Generative Pre-trained Transformer 3 (GPT-3) in this project. I also used the following tools to implement the user interface of the application and the logic that runs in the background: (i) Atlassian Forge; (ii) Cloud Functions for Firebase; (iii) Atlassian Design Components; and (iv) React. In addition to being able to access data about project teammates in Jira, we also needed to be able to retrieve contents and statistical information of Jira issues (requirements).

Challenges we ran into

AI tools were not technologies we were very familiar with at first, and our biggest challenges were to first understand how important it is to be able to input the right prompts to them in order to get correct (and useful) outputs, and then to try to intuit what is the "right" method. It was also a real challenge to use the issues in Jira to generate tales with the AI, while trying to make sure that the tale doesn't go out of its logical framework and get out of context. One of the things we've put a lot of thought into was enriching the tales with "illustrations" (also created by the AI) and making sure that they actually fit into the tale. Fortunately, we were spared from these challenges while realizing what we had envisioned on the Jira interface.

Accomplishments that we're proud of

The competition was full of debuts for me: This is my first long-running hackathon participation, and JiraTales is my first Atlassian add-on.

I am the ScrumMaster of a multidisciplinary team of talents from top schools, and have 15+ years of experience in software engineering. We have realized many successful IT projects by utilizing Atlassian tools: Automated billing systems, HR performance assessment systems, e-commerce plugins, data visualization tools, detection / positioning / sonar systems (in defense industry), and others are just a few of the many successful products that we have created.

Besides all this, it was a source of pride for me to be able to overcome the challenge of developing an AI-based gamification tool for Atlassian ecosystem after a relatively new entry into the AI universe.

What we learned

We realized once again how accurate the common assumption summarized as "garbage in, garbage out" actually is when we were trying to detect the "right" prompts into AI. Moreover, we first played the game we created amongst ourselves to see if it really increased our motivation and enjoyment of working on our project in Jira: The results were great. We were able to test how far JiraTales could take the tale (and whether it could do it so without losing context). Indeed, our desire to read its tale resulted in a faster and more enthusiastic completion of the assigned tasks. This fun work has led me to think that we can create further gamification add-ons for Jira (and for Atlassian tools in general) in the future.

What's next for JiraTales

First, we will consider enriching AI's tale parameters (theme, era, location, characters etc.) and giving Jira users different functionalities in path selection (at this point, I don't forget going ahead with the philosophy of 'less is more' since I believe it is crucial for greater app adoption). The feedback we will receive from users will be of great importance. In this regard, we will carefully select the features we think we can add. We anticipate that the ability to store the tales created by JiraTales in a library will be expected by users. We also want to continue to work on making the AI visuals that strengthen the tale even more relevant to the subject and context of the tale.