A(I)sk

Instant Insights in Jira: Just A(I)SK!

Comment

Inspiration

In the bustling corridors of modern workplaces, I frequently encountered a recurring theme: my coworkers often reached out with questions that were essentially "common knowledge," especially those sourced from our documentation. This observation wasn't a reflection of their expertise but rather the overwhelming nature of information in today's digital age. Despite having tools like Jira to manage projects, the sheer volume of data made it challenging to quickly find answers. That's when the idea of A(I)SK took root. Why not empower our Jira platform with the capability to instantly answer these common queries right within issue comments? A solution where instead of redirecting to lengthy documentation or waiting for someone's response, one could simply A(I)SK and get immediate clarity.

What It Does

A(I)SK revolutionizes the way teams interact with information within the Jira environment. Here's a breakdown of its capabilities:

  • Instant Answers: Users can instantly get answers to their queries. Instead of sifting through comments or diving into external documentation, just pose a question in a Jira issue comment and A(I)SK provides the answer.

  • OpenAI Integration: At its core, A(I)SK leverages the power of OpenAI. This ensures that the responses are not only pinpoint accurate but also contextually relevant.

  • Future Integration with Confluence: The horizon looks promising for A(I)SK. We're gearing up to integrate with platforms like Confluence. This will enable the AI to assimilate knowledge from existing documentation. What does this mean? Any user-centric guidelines, terminologies, or processes that are documented will be at the AI's fingertips. So, the next time a user has a detailed question about their documentation, A(I)SK will have the answer.

Our vision? To seamlessly bridge the gap between extensive documentation and real-time query resolution, making the entire Jira experience more intuitive and efficient.

Challenges We Ran Into

The journey of developing A(I)SK brought forth several challenges, especially when it came to teaching the AI about documentation:

  • Complexity of Documentation: Professional documentation often contains vast, intricate details. Grasping industry-specific jargon, multi-layered concepts, and the nuanced relationships between them proved challenging for the AI.

  • Contextual Understanding: Although OpenAI is a powerful tool, ensuring it truly understood the context of questions in relation to specific documentation was a hurdle. Recognizing words is one thing, but comprehending their contextual relevance is another.

  • Dynamic Nature of Documents: Documents evolve — they get updated, changed, and expanded. Keeping the AI aligned with the most recent version of the documentation became an ongoing task.

  • Avoiding Overfitting: Striking a balance was essential. We wanted the AI to provide broad answers when necessary but also zoom into the specifics of our documentation. Ensuring the model didn’t overfit to our documentation data was a challenge.

  • Feedback Loop Integration: For continuous improvement, we envisaged a mechanism where users could flag suboptimal AI responses. Building this feedback loop and ensuring its smooth integration posed its own set of challenges.

Navigating these challenges was instrumental in shaping A(I)SK into the robust tool it is today.

Accomplishments That We're Proud Of

One of the paramount accomplishments that stands out in our journey with A(I)SK is:

  • Leveraging ChatGPT: Harnessing the power of ChatGPT to answer user questions within Jira transformed the platform's interactivity. Not only did it streamline information access, but it also showcased the potential of infusing advanced AI models into everyday productivity tools, elevating user experience to unprecedented levels.

This integration became the cornerstone of A(I)SK, making it a beacon of innovation in project management tools.

What's Next for A(I)SK

The journey with A(I)SK has only just begun. As we look ahead, we're setting our sights on several exciting enhancements:

  • Parsing Documentation from Confluence: One of the most ambitious steps forward is enabling the AI to parse sample documentation directly from Confluence. This will empower A(I)SK to pull and understand data in real-time, providing answers that are not only accurate but deeply rooted in the specifics of an organization's documentation. This challenge, given the complexity and variability of documentation, is no small feat but promises a significant leap in utility.

  • Introducing Commands:

    • /remember: This command will facilitate users to instruct the AI to retain specific information, enhancing its knowledge base dynamically.
    • /link: Aimed at quickly retrieving links or references, this command will make sourcing documentation or relevant links a breeze.

While each of these steps brings its own set of challenges, especially the documentation parsing, we believe that they will collectively push A(I)SK to new horizons, redefining the boundaries of what's possible in AI-driven project management tools.

Built With

Share this project:

Updates