In this video, I walk through why n8n Chat Hub (chat v2) is a legit game changer for making workflows usable by non-technical people. Instead of forcing your team or clients to learn triggers, nodes, and how n8n executes, you can give them a ChatGPT-style interface where they can talk to your workflows. The part I’m most excited about is you can switch models mid-conversation (like GPT-5 to Claude Sonnet) and keep context, plus you can lock down who can do what for governance. I show exactly how to access Chat (you’ll likely need to update to the newer beta), how personal agents work (basically custom GPTs inside n8n), and how workflow agents let you call real automations from chat. We go through three examples: a super basic internal data-table lookup agent, a more advanced workflow that generates a video (and the key gotcha: you need an AI Agent at the very end with streaming enabled to return results to chat), and an image upload/binary file workflow using Gemini. I also cover publishing requirements, executions/cost considerations, and the new “chat user” role so people can use your automations without being able to touch your workflows.