The idea is:
I’d like to propose a product‑level feature for n8n:
On top of the existing AI nodes, AI Agents, HITL (human‑in‑the‑loop), Chat nodes, etc., add an embedded, first‑class AI workflow builder assistant (similar in spirit to Notion AI), and pair this with official web and desktop clients where this AI assistant is available out of the box.
In other words, instead of only exposing AI capabilities as nodes, n8n would also provide an “AI front door” for workflows: A place where a user can simply say “I want a workflow that does X”, and n8n’s own AI assembles the first version of that workflow, applying your Production AI Playbook patterns by default (HITL, approvals, logging, etc.).
My use case:
I’m a “non‑traditional developer but heavy user” of automation and AI workflows. In my day‑to‑day work I:
- Build and iterate many AI workflows and agents to process business data, respond to customers, and support content creation.
- Rely on many n8n integrations (databases, HTTP Request, Slack/Gmail, forms, AI nodes, etc.).
- But on busy days, starting from a blank canvas and manually deciding: trigger → nodes → tools → prompts → error handling → HITL, can feel like a high mental barrier.
A nice contrast is how Notion evolved after introducing Notion AI:
before, users had to understand and build everything manually;
after AI, many “I just want to achieve X” scenarios became: describe intent in natural language, let the product’s AI generate or modify the underlying structure.
In n8n, I’d love a similar path. For example, instead of manually wiring everything, I could say:
“Help me build a workflow: when a customer submits a form, use AI to draft a reply, send it to Slack for my approval, and after approval send it via Gmail.”
The assistant would then: - Propose or create the workflow skeleton (Trigger → AI Agent → HITL → Slack/Gmail → logging),
- Use your recommended patterns by default,
- Let me refine it via follow‑up prompts rather than configuring every parameter from scratch.
This would make me far more likely to open n8n and spin up a new automation even on very busy days.
I think it would be beneficial to add this because:
From my perspective, n8n already has a lot of powerful building blocks:
- AI Agent nodes, tool calling, HITL for tool calls, human approvals,
- Chat nodes, Form nodes, and a huge number of integrations,
- A strong Production AI Playbook with best practices for putting AI into production.
But from a new user / non‑developer perspective, the experience still feels quite “hand‑crafted”: you need a fairly deep understanding of nodes and workflow structure. This can lead to: - Potential users knowing n8n is powerful, but hesitating to adopt it because the initial learning curve feels steep.
- Existing users (like me) sometimes avoiding starting a new flow from a blank canvas when we’re too busy.
If there was an official AI workflow builder assistant, and your Production AI Playbook patterns were baked into its default behavior (for example: automatically suggesting HITL/approvals for high‑risk steps, proposing multi‑channel review templates, etc.), then: - New users could start by describing their scenario in plain language and get a workflow skeleton, then refine it through conversation.
- Existing users could edit workflows conversationally, e.g. “add a human approval step before this CRM write node”, “log all AI decisions into this Google Sheet”, “on timeout escalate to my manager via email”.
- Your best practices would move from static documentation into living product defaults and intelligent suggestions.
At the same time, if these capabilities were available through official web and desktop clients, users wouldn’t need to think about deployment or self‑hosting just to try n8n seriously. They could simply: - Install a desktop app or open a web app,
- Log in, and
- Immediately use the built‑in AI chat to create, edit, and run workflows.
This would make n8n feel like a daily “automation workbench” rather than “just another browser tab”, and would be especially friendly for non‑technical users and small teams.
Any resources to support this?
There are a few existing patterns that inspired this idea:
- Notion + Notion AI: Users can start from intent in natural language instead of fully understanding all blocks and database relations upfront.
- AI copilots in IDEs / document tools:
In VS Code, AI assists with writing and refactoring code.
In Notion, AI assists with generating and transforming content.
In n8n, this could become a “workflow copilot” that helps generate, explain, and adjust workflow structures.
Also, the patterns already described in n8n’s Production AI Playbook (e.g. where to add HITL, how to handle high‑risk actions, how to structure approvals and logging) seem like a perfect fit to be turned into default suggestions and behaviors inside such an assistant.
Are you willing to work on this?
I’m not a backend or systems engineer, but as a heavy user I’m very happy to:
- Share more detailed real‑world use cases and requirements from my day‑to‑day workflows,
- Participate in testing prototypes / beta versions of such an assistant or client,
- Provide feedback from the perspective of a non‑traditional developer who builds a lot of AI workflows and cares about production‑grade reliability.