Skip to content
Aakash GuptaAakash Gupta

Seriously, Please Watch This Before You Declare n8n Dead

n8n was the hottest AI tool in the world in 2025. Then Claude Code and Cowork arrived and people started declaring it dead. Jan Oberhauser, founder and CEO of n8n, walks through what his product does that a Claude Code agent can't, with a live demo of reliability, auditability and human-in-the-loop approvals. Full Writeup: https://www.news.aakashg.com/p/n8n-vs-claude-code Transcript: https://www.aakashg.com/claude-code-killed-n8n-its-ceo-disagrees/ Timestamps 00:00 - Intro 02:28 - Do you still need n8n in the age of Claude Code 07:09 - Ads 10:17 - Did Claude Cowork and Claude Code hurt n8n 11:38 - n8n's revenue, users and enterprise customers 14:44 - Demo, what n8n does that a Claude Code agent can't 22:46 - What reliability and auditability actually mean 24:47 - Ads 29:40 - Handing a workflow to a teammate, and where n8n shines 32:21 - What PMs should build after their first agent 41:44 - Why sprinkling AI on top only gets you 10 to 30% 44:23 - Killing the lead gen target and per-seat pricing 54:05 - How n8n builds product and hires AI PMs 🏆 Thanks to our sponsors 1. Customer.io: Send smarter messages using your product data - http://customer.io/productgrowth 2. Kameleoon: Leading AI experimentation platform - http://www.kameleoon.com/prompt 3. Bolt: Ship AI-powered products 10x faster - https://bolt.new/solutions/product-manager?utm_source=Promoted&utm_medium=email&utm_campaign=aakash-product-growth 4. Jira Product Discovery: Plan with purpose, ship with confidence - https://www.atlassian.com/software/jira/product-discovery 5. Product Faculty: Get $150 off their #1 AI PM Certification via code AAKASH150 - https://www.productfaculty.com/?code=AAKASH150 Key Takeaways 1. n8n and Claude Code are different products, and you need both - Claude Code runs Anthropic models in your terminal as an agentic tool. n8n is the orchestration layer that connects your tools, models and data sources on a visual canvas. 2. Reliability means fallback models, self-hosting and maintained tooling - Every provider goes down, so you set a fallback model in the same workflow. You can self-host n8n next to your own data. Each integration is code n8n writes, tests and maintains, so an API change gets fixed once for everybody instead of by every builder separately. 3. Auditability is the thing code can't give you - With generated code you see the input and the output, not the decisions in between. n8n shows every past execution step by step, with the data going in and out of each node, and lets you rerun half a workflow from a specific data point. 4. AI plus deterministic logic plus human in the loop - AI is one tool, not the whole solution. An if statement is cheaper, faster and 100% reliable, so the two belong together. 5. The AI assistant builds the workflow for you - Jan prompts it the way you would prompt Claude Code. It asks clarifying questions, thinks for about eight minutes, and returns a working multi-tool agent. Extending it afterwards with one-on-one scheduling and Google Contacts took another five minutes. 6. Start small and skip AI when you don't need it - Companies that try to transform everything at once spend weeks building the wrong thing. The simple wins are also the ones that get your colleagues interested. 7. Where n8n uniquely wins is anything business-critical - Security orchestration, compliance, employee onboarding and offboarding, DevOps, monitoring. The rule of thumb Jan gives is that the more reliability and security matter, the more n8n shines, and that covers more or less anything that isn't a personal use case. 8. Sprinkling AI on top gets you 10 to 30%, being in the value chain gets you 10x - Adding an AI button somewhere is not a strategy. n8n's bet was that when someone decides to build an agent, they build it in n8n. That choice is what produced 10x growth in a year. 9. Deleting the metrics that make money faster - No lead gen target and no per-seat pricing, because profitability means n8n doesn't need investor money and can think long term. The company also doesn't push free self-hosted users onto paid hosting. The internal goal moved from a billion in ARR to a billion users, so nobody mistakes the mission for money. 10. Talent density over headcount, and technical PMs over polished ones - The target is a billion users with fewer than a thousand employees, so the hiring bar stays high. Jan wants tinkerers who run home automation and understand scale, evals and reliability. Live problem-solving sessions beat take-home tasks now that AI can do the take-home for you. đŸ‘šâ€đŸ’» Where to find Jan LinkedIn: https://www.linkedin.com/in/janoberhauser/ X: https://x.com/JanOberhauser đŸ‘šâ€đŸ’» Where to find Aakash X: https://x.com/aakashgupta LinkedIn: https://www.linkedin.com/in/aagupta/ Newsletter: https://news.aakashg.com 🧠 About Product Growth The world's largest podcast focused solely on product + growth, with over 200K+ listeners. 🔔 Subscribe and turn on notifications.

Jan OberhauserguestAakash Guptahost
Oct 4, 20261h 9mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Why n8n still wins for reliable, auditable AI workflows

  1. Jan Oberhauser explains why n8n remains relevant alongside Claude Code by positioning it as an orchestration layer for production-grade, business-critical automation rather than an agentic coding environment.
  2. The episode defines n8n’s differentiators as reliability (self-hosting, tested integrations, retries, model fallbacks), auditability (node-level execution logs and replay), and governance (human-in-the-loop approvals).
  3. Jan shares growth signals—including $100M+ ARR, ~1.5M active users, 1,200+ enterprise customers, and 10x revenue growth—arguing that ‘n8n is dead’ claims reflect hype-cycle dynamics, not fundamentals.
  4. A live demo shows how n8n’s AI assistant can generate and iteratively extend workflows (e.g., Gmail/Calendar agent) while keeping the underlying system visible and editable for teams.
  5. The conversation covers product and org strategy: small squads, preference for highly technical “builder” PMs, dedicated evals ownership via an AI Trust team, and an efficiency goal of massive scale with comparatively small headcount.

IDEAS WORTH REMEMBERING

5 ideas

Claude Code and n8n solve different phases of the problem: prototype fast vs run reliably.

Jan argues that agentic coding tools are great for rapid prototyping, but n8n becomes essential once a workflow must run consistently, be inspectable, and be safe to operate in production. He frames n8n as the orchestration layer that connects apps, data sources, and multiple LLMs with explicit control points.

Reliability isn’t just uptime—it’s controlled execution, safe fallbacks, and maintainable integrations.

Reliability includes model fallbacks (e.g., switching if a provider is down), deterministic logic (ifs/branches) where AI isn’t needed, retries/operations tooling, and the ability to run on your own infrastructure. n8n’s maintained integrations also reduce fragility when third-party APIs change.

Auditability is a first-class feature: you can see what happened, why, and where it broke.

Auditability in n8n means you can inspect the workflow structure and also replay/inspect historical executions node-by-node, including what data went in/out and which tools were called. This is positioned as critical for compliance and for debugging AI-driven decisions versus treating the agent like a black box.

Human approval steps are a practical safety layer for agentic workflows.

The demo highlights approval gates (“human-in-the-loop”) that prevent an agent from taking high-risk actions (like sending an email or creating a calendar event) without confirmation. This blends AI autonomy with explicit governance and reduces the blast radius of hallucinations or prompt drift.

Despite ‘n8n is dead’ narratives, the company reports major scale and acceleration.

Jan claims n8n surpassed $100M ARR and is “way across” that now, with ~1.5M active users, 200k+ GitHub stars, and 1,200+ enterprise customers on the enterprise solution. He also states n8n grew revenue 10x over the past year and that even frontier labs use it for business-critical workflows.

WORDS WORTH SAVING

5 quotes

I think we have been called dead probably a thousand times. I don't remember half of the things that, that killed us over the years.

— Jan Oberhauser

Like, Claude Code, it generates literally ten thousand lines of code that nobody can inspect and nobody knows if, if, if it's actually doing the right thing.

— Jan Oberhauser

You cannot just, uh, rely on it working ninety-five percent of the times, but you c- have to be a hundred percent sure that it really works.

— Jan Oberhauser

That's literally our, our mission is to try to give everybody who uses a computer and tech superpowers, and that's what, what we're thriving for, um, these days.

— Jan Oberhauser

Because of the way n8n is built, like we are actually sustainable. Like right now, we, we are actually creating a profit.

— Jan Oberhauser

n8n vs Claude Code vs Claude Cowork positioningReliability: self-hosting, retries, maintained connectors, model fallbacksAuditability: execution traces, inspection, debugging, replayHuman-in-the-loop approvals and governanceEnterprise adoption metrics and growth claimsTemplates and common use cases (scraping, monitoring, security ops)Product org design, technical PM expectations, evals/AI Trust team

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.