January 2, 2025

In this blog post, we will discuss strong points of JQL but also use cases and a reasoning for an enterprise search and retrieval augmented generation with Atlassian Jira.

The author of this blog post believes that Atlassian Jira is one of the best project management and issue tracking tools. It is well-suited for a wide range of projects - not only for software development or IT topics but also beyond. And it scales with team sizes from a few team members to dozens and more.

Atlassian Jira comes with a strong embedded, advanced search which is powered by its own query language, Jira Query Language (JQL).

So why is it a good idea to still index Jira projects into an enterprise search?

Successful Atlassian Jira Usage and Knowledge Management

For a successful Jira usage it is crucial that all team members understand the ticket processes. All team members must also document their work in a reasonable way. For instance, when it comes to project work, a general rule for all RheinInsights members is the following.

People who are not deeply involved in a project must be able to get a quick understanding of the project status through the Jira dashboards. If one dives into a ticket, the current state, the scope must easily become clear (without asking involved persons). If an issue was solved, the solution must be documented and how this ticket ended. For instance, if code went productive, it must be documented when and in which version.

Why are we writing this? Because a great documentation standards in Jira lie the foundation for an even better knowledge retrieval.

Jira Query Language and Search Use Cases

Jira offers its own search functionality. It is well-suited for retrieving issue, if you know what you are looking for, as well as for project reporting. It has a rich query syntax which and when used as an advanced search reminds strongly on SQL syntax.

However, if it comes for getting quick answers, overviews or retrieving issues where you do not know the exact wording, JQL quickly reaches its limits. For instance this generally holds true for the following, common use cases

1. Finding most relevant results

2. Generating answers to problems which have been solved in the past

3. Finding information which stand in a certain, broader context

4. Retrieving a specific issue without knowing the exact wording which was used in the ticket

5. Getting a quick overview of / browsing many search results

Enterprise Search with Jira Documents

As written above, the built-in search is perfect for quickly retrieving issues which are not too much in the past. Or it works well when you know the exact wording and their existence.

Besides the built-in search, you can search through your Jira issues, comments and attachments after indexing these in your enterprise search or your secure vector search.
In turn, your enterprise search offers

• A full-text search for all of your Jira data

• State-of-the-art linguistics such as synonyms

• For Keyword search: lemmatization and tokenization

• Vector search (cf. e.g. RheinInsights - Documentation - Search Engines)

• Lean search result presentation with hit-highlighting and teasers

• Great search result relevance

• Secure search with early binding security trimming (see e.g. RheinInsights - Blog - Permission-Based Retrieval Augmented Generation)

Beyond Enterprise Search - Sourcing fast Answers Based on Your Knowledge

Indexing your Jira knowledge in a search engine is the foundation for offering great answers to your employees and partners. Afterwards, you can integrate your Jira documents into your (secure) RAG-application, for instance your Q&A-bot.

Such a bot application can be used in the following common use cases

• Case deflection: offers answers based on ticket solutions in a specific Jira project so that users can solve problems on their own

• Service support: searching and summarizing issues, in particular for agents in the field

• Retrieving experts based on their Jira interactions

• Summarizing assigned and open issues

• Retrieve answers on more broadly formulated questions

Fulfillments and taking actions in Jira. Please note that through the Jira APIs it is also possible to implement actions from the bot against Jira. For instance, the bot can open a new issue in Jira. Here you would extend the bot to carry out Jira-actions on behalf of the user through the REST APIs, Jira offers. In contrast, enterprise search connectors will only allow for crawling and indexing Jira issues.

Getting Started - The RheinInsights Retrieval Suite

Is it complicated to integrate Atlassian Jira in your enterprise search or vector search? Actually, it is not.

Our RheinInsights Retrieval Suite offers a connector for Atlassian Jira Cloud and also one for the on-premises editions Atlassian Jira Data Center and Standalone. To get started, do the following

1. Download our RheinInsights Retrieval Suite from Downloads.

2. Quickly perform the base setup

3. Configure the according Jira connector

   1. Configure a connection to your Jira instance

   2. Configure a connection to your favorite search engine, for instance Microsoft Azure AI Search or Microsoft Search

   3. Configure the content processing pipeline, in particular if you want to use secure vector search.

   4. Crawl your Jira instance

4. Integrate the search results into your search experience

   1. Either use our built-in enterprise search experience (cf. RheinInsights Search Interface)

   2. Integrate the Jira results into your existing search or bot experience

If you are interested in a trial of our Suite, then please reach out to us: Contact.