n8n

When we started planning our first-ever conference, I never imagined we would have this lineup on the agenda: SAP, NVIDIA, Microsoft, Accenture, Capgemini and Boston Consulting Group (BCG) I am incredibly grateful to these teams for joining us at "In The Loop: and sharing how they are using n8n to tackle some of the biggest challenges around AI and automation. We will hear about: - AI sovereignty  - Scaling n8n across a large enterprise like SAP - How n8n will be embedded in Joule Studio - Choosing the right workflows to maximize ROI And we are only getting started.  I would love to see you in Berlin on October 14 for our first In The Loop. https://n8n.io/intheloop/

We had a great few days at HumanX Europe in Amsterdam 👏 Thanks to everyone who stopped by the n8n booth, shared what you’re building, and talked to us about the challenges you’re running into. If we didn’t get a chance to chat in Amsterdam, drop us a comment or DM. We’d love to hear from you!

Jamie Madden, our Senior Education DevRel at n8n, will be speaking at NVIDIA GTC in Berlin this October. His session, “Your Agent Doesn’t Get Root: Governing AI Actions With n8n and NVIDIA NemoClaw,” will cover practical ways to let AI agents take action without giving them broad permissions. He’ll walk through sandboxed agents, MCP tools, human approval steps, network policies, and model selection, all with a focus on keeping agent actions controlled and secure. If you’re at GTC, join the session and come say hi to us at booth #3029. 👋 🎟️ Register here: https://lnkd.in/eXhMB429

Moving an AI agent from prototype to production takes more than a working prompt. You need guardrails, reliable tool execution, controlled context, evaluations, tracing, versioning, and a rollback plan. Our new article breaks the process down into 6 practical stages — with examples of how to implement each one in n8n:

Agentic automation only scales once identity, permissions and a reusable pattern are in place. Oliver Dörr (Microsoft) and Raphael Gomez (n8n) take the stage at In The Loop on 14 October to walk through how that works in practice. They’ll discuss → What Microsoft built with Agent 365 and why agents needed an identity of their own → Where n8n fits across the Microsoft stack as the agent runtime → What the two teams are building next. If you're running automation on Microsoft, this one is worth the slot. Link to registration in the comments. 💬 #InTheLoop2026

Heading to Bits & Pretzels next week? We're co-hosting the OpenAI Builder Lounge in Munich on Sept 29. Our Founder and CEO, Jan, is doing a fireside chat on how we think about AI workflows, automation, and giving builders more control over what they build. Spots are limited, so don't sleep on it. Link in the comments. See you in Munich! 🥨

This week’s featured workflow runs on scheduled weekly and monthly triggers to generate unified marketing performance reports. It processes multiple websites by collecting analytics data, paid ads performance, and CRM leads, then calculates KPIs and insights automatically. Built by our Ambassador Avkash Kakdiya, it connects Google Analytics, Google Ads, Meta Ads, and HubSpot, then sends the finished report straight to your inbox and keeps a history in Google Sheets. If you're managing marketing across several websites or clients, this saves you from rebuilding the same report by hand every week: https://lnkd.in/eMum3uVw Thanks for sharing, Avkash! 👏

Deploying an AI agent is easy. Deploying one you can actually trust in production is harder. Our latest guide covers the controls that matter when moving agents into live use — from least-privilege tools and human approvals to execution limits, monitoring, evaluations, versioning, and progressive rollouts. Plus, practical examples of how to implement these safeguards in n8n: