Scaling n8n inside regulated enterprises: 1M executions, 3,000 hours back, zero engineering bottleneck

In a regulated company, internal automation loses twice. Compliance and audit slow every change to a crawl. And engineering, correctly, puts the product first — so the automation backlog never moves. This talk is about two Brazilian companies that got past both, with n8n in production today. Anti-fraud technology — Certta. A verification hub serving 250+ clients, handling 200 production change requests a month, where a broken workflow means a fraud gets through. Instead of queueing automation behind the product roadmap, Certta built an automation Center of Excellence on n8n and pushed building out to the business itself: Engineering, HR, Finance, CyberSec, Support, Marketing and Legal now build their own workflows under shared governance. Over 1M executions, processes 75% faster on average, 3,000 hours returned to the teams. Insurance — Icatu Seguros. A.V.I, an n8n-orchestrated WhatsApp assistant that lets 1,000+ insurance brokers run their whole portfolio, not just quote. From idea to production in seven weeks, quotation time down from about five minutes to under 40 seconds, and second place at Gartner's Eye on Innovation Awards for Insurance 2025. From idea to production in seven weeks, quotation time down from about five minutes to under 40 seconds, and second place at Gartner's Eye on Innovation Awards for Insurance 2025. We built it before n8n's Agent framework existed — what took ten nodes then would be three today, and we'll show both. We will show the real architectures, the governance model that got past auditors and security teams, and the CoE patterns that let non-engineers build without every team improvising its own. We will also tell you the mistake we made first, why it stalled, and what unlocked it. Presented by the team at ed, n8n's Expert Partner Premium in Brazil, who built both.