As organizations grow, so are their automation processes and technical debt caused constraints. Triggers and classes are mixing with Flows, Process Builder, Workflows and over time any change and improvement becomes too complex, too hard or even impossible. While we all try to write clean code and follow best practices, we can’t ignore the complexity that undeniably keeps growing, especially in highly customized orgs with a lot of integration points. In our session we will present a case study of automation refactoring in Service Cloud of Okta’s Salesforce org. Since Okta, as identity management provider, has high and very strict SLAs, customer support process is complex and requires extensive automation in order to ensure support agents’ productivity and timely response. We started with a deep technical debt on Apex side, dozens of Workflows, Flows and Processes and we re-built an automation keeping in mind that we must improve the process and that our code will be legacy code for future developers and admins working in this org. In our session we will show how we made changes to make code more readable and understandable and our declarative processes streamlined and enabled for easier maintenance and improvements based on business needs. We will also present and compare runtimes for before and after processes explaining why the changes we made are improving performance and how that impacts the whole system, not just subgroup of Salesforce users in Service Cloud. As a special treat we will provide a list of gotchas that we learned along the way and highlight edges cases that has to be considered in similar projects. Audience will leave the session better prepared to dive into their own automation refactoring projects and ready to follow best practices that ensure testability, maintainability and extensibility of the code.