I am a software engineer with more than 10 years of experience and an obsession with optimizing processes and making life easier both in personal routines and in large projects.
After reading about atomic habits, I adopted a professional creed focused on optimizing everything, from daily tasks to complex deliverables, applying that mindset to software quality and QA team leadership.
As QA Manager of a team of 7 testers, I always believed that process optimization is the pillar of quality assurance. By combining my management experience with solutions based on artificial intelligence and multi-stage agents, we built workflows that drastically improved team efficiency.
Main objective: eliminate manual and repetitive work in test case creation. Setting up integrations with APIs like Jira's can be tedious and full of technical details, but the reward is worth it when we manage to automate and standardize processes.
We built an AI agent that transforms Jira tickets into structured test cases ready to publish in minutes. The flow accepts a ticket ID like TEST-123, retrieves the summary, description, and acceptance criteria, filters the Jira JSON to extract only the essentials, and generates cases with Title, Preconditions, Steps to reproduce, and Expected results.
The agent also proposes edits, adds negative cases, and merges before publishing. With direct integration to the test management tool Testomat.io, the team can review in 3 to 5 minutes and publish with one click, compared to the 45 to 60 minutes the manual process took on complex tickets.
Key technical components: advanced GPT model for language generation, access to the Jira API for ticket retrieval, and push to Testomat.io to publish cases. Complementary tools included format validation and exception handling for missing fields or unclear acceptance criteria.
Main challenges: OAuth 2.0 configuration and adjusting callback URLs, API permissions and scopes, the verbose Jira JSON that requires filtering, and managing edge cases to avoid erroneous outputs. Patience with the integration is worth the productivity result.
Benefits for the team: reduced time in test case creation, standardized formats, greater coverage with inclusion of negative and edge tests, and elimination of copy-paste fatigue thanks to direct integration. In numbers, efficiency improved and work consistency increased significantly.
Lessons learned: APIs are the glue of integrations, clarity in prompts determines output quality, and publishing directly from the agent changes the game. Adding project documentation to the agent's knowledge avoids generic or out-of-context tests.
At Q2BSTUDIO, a company specialized in software development, custom applications, and custom software, we implement similar solutions for clients that require artificial intelligence applied to the software lifecycle. We offer artificial intelligence services, AI for businesses, and AI agents that automate repetitive tasks, as well as cybersecurity to protect integrations and sensitive data.
Our services include cloud architectures with AWS and Azure cloud services, business intelligence and reporting services with Power BI to turn data into actionable decisions. We are specialists in custom applications that integrate artificial intelligence, cybersecurity, and cloud services to deliver complete and secure solutions.
Conclusion: AI is not here to replace QA engineers, it is here to enhance their work, making them faster, more consistent, and focused on what adds the most value. If you lead a QA team or are an engineer, it is worth exploring AI agents and API workflows to automate test case generation and increase software quality.
We invite you to contact Q2BSTUDIO to design your own AI agent, integrate your tools like Jira and Testomat.io, and deploy custom software solutions, artificial intelligence, cybersecurity, AWS and Azure cloud services, business intelligence services, AI for businesses, AI agents, and Power BI that drive your results.
And now your turn: have you tried connecting a GPT agent with Jira or your test management tool, and what challenges did you encounter doing so?
