Dylan Ratcliffe, Overmind
How much precious engineering time are you spending on changes that don’t really need it? In this talk, we’ll reveal how we analysed nearly 37,000 infrastructure modifications to rigorously quantify "toil"; the work that slows teams down without adding real value. We’ll share the statistical techniques and practical models we developed to find the best opportunities for safe auto-approval, and show how you can use similar methods to identify (and eliminate) wasted effort in your own workflows. We’ll also discuss pitfalls, trade-offs, and how AI can take your approval process even further. Whether you’re a platform engineer or an SRE, you’ll learn how to free up your team to focus on what really matters while keeping your systems safe and reliable.
Dylan Ratcliffe is the Founder & CEO of Overmind, where he leads efforts to harness AI for preventing outages and giving platform teams real confidence in deploying infrastructure changes. Before founding Overmind, Dylan spent nearly seven years at Puppet, holding senior engineering and leadership roles across Australia and the UK, including Senior Manager of Professional Services for EMEA. At Overmind, he has raised funding from some of Silicon Valley’s most successful venture capitalists, fueling his mission to revolutionise change management and deployment safety. Based between London and San Francisco, Dylan is also passionate about motorcycle racing, various electronics projects, and travel.
author = {Dylan Ratcliffe},
title = {How We Used Statistics to Find Toil among 36,000 Changes},
year = {2025},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}