Profiling helped fix slowness in Sentry's AI Autofix by identifying two bottlenecks: redundant repo-initialization and unnecessary thread waits during insight generation. The analysis was made possible by Sentry Profiling, which captures data about function calls, execution time, and resource usage to identify performance bottlenecks at the CPU & browser level that are impacting real users. By dogfooding their own profiling tools, Sentry fixed a problem that saved tens of seconds off each user interaction with their AI agent, providing a four orders of magnitude bigger impact than those "milliseconds matter" approaches. The fixes involved caching and thread optimization, reducing execution times by tens of seconds and making Autofix noticeably faster and more responsive without requiring any architectural changes.