Challenge
A US-based fintech and B2B payments company was burning senior engineering hours on FTP support tickets that looked simple but never were. With 15 to 25 of these tickets coming in every month, each one triggered a multi-step investigation across Splunk, Rundeck, internal databases, and Jira: was it a wrong environment? Banned IP? Locked account? Bad credentials? Missing ticket data? The EOPS team’s average triage time exceeded one hour per incident, and the cost only scaled with customer growth.
Solution
We built a modular automation system on n8n, nine sub-workflows in total, that turns every FTP ticket into a fully orchestrated diagnostic workflow. Claude parses raw Jira tickets into structured data. If anything’s missing, Splunk fills the gaps automatically. Every diagnostic check (connectivity, account status, IP ban analysis) runs in parallel, and results write back to Jira as a structured ADF summary. Known failure modes like IP bans or locked accounts trigger safe automated remediation. Only genuinely ambiguous cases escalate, and those go to Slack with the exact missing context spelled out.
Business Value
Triage time dropped from 60+ minutes to under 60 seconds. The 15 to 25 FTP tickets coming in monthly now triage themselves automatically end-to-end, with zero manual effort needed for known failure scenarios. Delivered in 3 months by a 2-person IWConnect team, the system ships as 9 modular sub-workflows that the SRE team can extend independently. Engineering capacity moved to higher-value reliability work and the next round of automation projects.
Download the full case study to see the architecture, sub-workflow design, and AI prompting decisions behind the build.
