Smart Document AI

Our Atlassian Forge-powered AI app makes Confluence so easy that even a child would love to use it. We have redesigned the way people get their answers from confluence. Transforming data to insights.

Comment

Inspiration

We are using Confluence in our organization. It contains all our company standards and best practices for each role from developers to managers or stakeholders. It’s a bible for everything yet almost 90% of people are not using it, treating it like their school textbooks.

There was always a need for efficient, data-driven insights in the modern workplace. As businesses increasingly rely on digital documentation and collaboration platforms, such as Atlassian's Confluence, the potential for extracting meaningful information from this wealth of data becomes evident. Our team recognized the opportunity to empower the Atlassian community with a solution that goes beyond traditional document management and that gave rise to our Product ‘Smart Document AI’

What it does

The Smart Document AI empowers users to extract insights from their own Confluence, offering data-driven recommendations and answers within the Atlassian ecosystem. The app enhances document management, answering questions and unlocking valuable knowledge from content, all with a simple 1 click setup.

How we built it

  1. Ideation: We started by analysing problems in the current ecosystem and identifying gaps.

  2. Prototyping: After analysis, we started prototyping our Confluence Solution and we realised this is something that can be used across the community.

  3. Implementation: Using all our knowledge of AI, AWS and the forge platform we started working on the solution and were able to create a simple yet powerful AI-powered Confluence solution.

Challenges we ran into

  1. Data Size: Confluence works as the main repository for Organisations so Managing 1000s of pages, requires adapting to a vector-based memory for LLM Models.

  2. Security and Privacy: Safeguarding sensitive information within the documentation and ensuring data privacy presented complex challenges, which we addressed diligently with enterprise Architecture and AWS Safety and Security protocols.

Accomplishments that we're proud of

We are very new to the Atlassian forge platform, before Hackathon we never worked on it but within a few weeks of using Atlassian forge we understood its power and how simple and convenient it is to test and deploy any application on the forge platform. We had ideas but with Atlassian Forge, we got a whole new domain to widespread our ideas into solutions.

What we learned

  1. AI Integration: We deepened our understanding of AI integration into real-world applications, using advanced techniques like large language models and index-based memory for storing thousands of pages to provide valuable insights.

  2. Atlassian Forge: This is the first time we explored the capabilities of Atlassian Forge, allowing us to seamlessly deploy and integrate our app into the Atlassian ecosystem. This experience opened up new possibilities for enhancing collaboration and productivity.

  3. User-Centric Design: Building a user-friendly interface and ensuring a smooth user experience taught us the importance of putting the user at the centre of our design and development efforts and making the onboarding experience as smooth as possible to attract customers.

What's next for Smart Document AI

  1. Enhanced AI Capabilities: We'll continue to advance our AI models for even more accurate document analysis and insights, potentially incorporating more advanced LLM Models and Vector Management.

  2. Integration with More Platforms: We have already done a POC on integrating our solution with Slack, a feature currently in the process of being optimized for widespread use. The synergy of Slack or any other workplace communication channel with Confluence, supercharged by AI, elevates the utility of this application, empowering organizations to consistently utilize their invaluable

Share this project:

Updates