Summary
Challenge: The incident management process was highly inefficient, requiring human agents to manually review around 2,000 lines of ScriptRunner data per error log. This task typically took up to four hours to complete, making it time-consuming and prone to delays.
Solution: An AI agent was developed in Python using AWS Strands to automate the process. The agent is able to quickly and accurately parse error logs from CSV files and match each ScriptRunner Listener error to the correct script, integrating seamlessly into the existing codebase.
Outcome: The implementation led to approximately 75% time savings for human agents, significantly reducing manual effort. It also streamlined the workflow by automating a previously labor-intensive task, resulting in faster and more efficient incident resolution times.
Background
Discover how we embedded an AWS Strands AI Agent to tackle inefficient incident management challenges and transformed our client’s approach to Scriptrunner data handling.
Our client, an international insurance firm, was struggling to manually process error logs when managing incidents. The team were handling thousands of lines of data, leading to inefficiencies and slower resolution rates. To solve this, we created a custom AI Agent using AWS Strands.
AWS Strands is an open-source Software Developer Kit (SDK), which enables teams to build agentic AI capabilities in Python or TypeScript, as part of their code-flow.
Challenge
When an incident occurs, such as a system outage, Scriptrunner Listeners generate thousands of lines of errors. This meant that our client’s team were having to manually comb through each line to identify which Listener was producing the error. To further complicate matters, each Listener has its own ID – a convoluted jumble of letters and numbers. So team members needed to determine both the error and the relevant Listener ID, which took further investigation.
For an average error log with ~2000 lines, this work used to take a human agent ~4 hours to complete. This required huge manual effort, and led to delays in incident resolution.
Solution
To solve this, we created an AWS Strands Incident Management AI agent, written in Python. The Agent can now parse thousands of error lines in a .CSV file, quickly and accurately. Based on a mapping table between script name and ID, it can efficiently match the correct issue key to the correct script. This is a real game-changer for our client’s incident management processes.
Benefits
Automating this process with an AWS Strands AI agent has delivered time-savings of ~75%, streamlining a previously manual and time-consuming process.
We further developed this solution by adding an automatic data fix to one of the scripts. This has also improved incident resolution rates, as we were able to use our existing code, in line with the new agent, to apply a fix.
Ultimately, our new AI Incident Management Agent, built with Python in AWS Strands, has delivered significant time-savings, reduced human intervention, and enhanced issue resolution.





