Natural language to JQL

Having trouble finding tasks? Tell us what you're looking for
Give us a moment to understand your query
Go straight to the search results or copy your ready JQL query
Save your favourite translations for later!

Inspiration

The idea came from a desire to help business-oriented people wanting to facilitate their decisions with data provided by Jira. Often times, these people don't have the technical background, which is essential to unleashing Jira's full potential. Right then, we realized that it's necessary to bridge that gap and lower the entry barrier for non-technical Jira users.

We've embraced the latest developments of technology and came up with a solution to adress this problem. We’ve reached the point at which we decided it's universal enough to be shared with other companies where decision-making parties may also be in need for a quicker way to navigate through Jira issues.

What it does

"Natural language to JQL" converts user's text input in natural language and translates it to a JQL query.

Once the input is translated, by clicking a button, user can apply the query and instantly see the results in the issue navigator. Also, the query can be copied to clipboard and used in every place where JQL is applicable. For more comfortable experience and better accessibility, the app supports both keyboard and speech-to-text user inputs, written or spoken in multiple languages.

How we built it

We used a machine learning algorithm to refine the keywords from the natural language input and then translate them to an optimized JQL query. For the speech-to-text feature, we utilized the native API, provided by the world’s most popular web browsers. Apart from that, the Forge platform enabled us to implement the remaining logic of the app such as redirecting to the search results or quick inline editing of the JQL query. All that was accomplished in a fraction of the time we’d normally spend on building it the conventional way.

Challenges we ran into

The biggest challenge was posed by the algorithm itself, which sometimes struggled with accuracy, especially when dealing with many filling words in the input. It required us to adjust our original designs and let users edit the translation manually. Fortunately it didn't turn out to be a critical problem, and most importantly, we managed to work our way around it. We are happy with the knowledge and experience gained along the way.

Accomplishments that we're proud of

We firmly believe that many business-oriented users and decision makers around the world are facing similar technical barrier when it comes to handling Jira. That’s why creating a tool that has the potential to improve their work is the accomplishment we’re the most proud of. We also have a strong feeling that this solution might be a stepping stone towards the future with more accessible technology.

What we learned

The most important takeaway from our journey was simple yet very often overlooked by others, especially in the time of apps overloaded with features. It’s vital to remember that creating an app for Jira is all about finding a working solution to a real problem.

What's next for Natural question to JQL

This is the kind of solution that we want to see blossom in the future and continuously serve its purpose to its users. We're hoping to share it with other customers, through the Atlassian Marketplace, along with some exciting functionalities:

Integrating the algorithm into the native issue navigator searcher for even more seamless experience.

Performing other actions inside Jira such as bulk operations or creating issues, based off of input in natural language.