Aakash GuptaSeriously, Please Watch This Before You Declare n8n Dead
CHAPTERS
- 0:00 – 3:19
Is n8n actually “dead”? Hype cycles, Google Trends, and why orchestration still matters
Jan responds to the viral “n8n is irrelevant” narrative and explains why Google Trends dips often reflect hype-cycle normalization, not product death. He positions n8n and Claude Code as complementary: Claude Code for rapid prototyping, n8n for operationalizing workflows that must be dependable in real businesses.
- •n8n has been “declared dead” many times; hype cycles create predictable trend arcs
- •Claude Code vs n8n: terminal-based agent building vs orchestration layer for systems
- •Business-critical automation requires reliability, security, and visibility—beyond a prototype
- •n8n used alongside Claude Code: prototype fast, then migrate to n8n for production
- •Self-hosting and model/tool/data connectivity are core n8n differentiators
- 3:19 – 7:07
Why n8n persists in the agent era: reliability, security, auditability, and model flexibility
Jan details why n8n remains valuable even as agentic coding tools improve. He highlights model flexibility (switching/routing across providers), self-hostability for privacy, and the ability to connect multiple models and data sources into secure, inspectable workflows.
- •Model flexibility: swap models as capabilities and pricing change daily
- •Multi-model orchestration: use OpenAI/Anthropic/open-source models together per task
- •Self-hosting and data privacy increasingly drive tool choice
- •Visual workflows reduce black-box risk versus large generated codebases
- •n8n adoption includes even frontier labs and compliance/security use cases
- 7:07 – 9:58
Ad break: Customer.io and Kameleoon (prompt-based experimentation)
Aakash shares sponsor messages focused on behavioral messaging automation (Customer.io) and reducing developer involvement in experimentation (Kameleoon).
- •Customer.io: event-driven messaging, AI agent to build journeys, MCP server integration
- •Examples of improved onboarding and open-rate/conversion uplift
- •Kameleoon: prompt-based experimentation to generate variants quickly
- •Targeting cohorts and KPIs without heavy engineering time
- 9:58 – 11:39
Did Claude Cowork/Code hurt n8n? Shifting from personal automations to business-critical workflows
Aakash asks directly whether new agent tools impacted n8n. Jan argues these launches increased awareness and changed the mix of use cases: away from novelty/personal tasks and toward mission-critical workflows where governance and correctness matter.
- •New agent builders often create spikes in attention and opportunity, not just competition
- •n8n’s biggest shift: more business-critical, less “personal toy” automations
- •Mission-critical workflows require 100% confidence, not 95% success rates
- •Market hype can mislead; real adoption is reflected in usage and ROI
- 11:39 – 15:20
n8n’s scale metrics: ARR, active users, enterprise customers, and SAP distribution
The conversation turns to company performance and enterprise traction. Jan confirms n8n is above $100M ARR, shares active user and enterprise-customer counts, and explains the significance of SAP embedding n8n for its customer base.
- •Jan confirms n8n is ‘way across’ $100M ARR and growing strongly
- •~1.5M active users; ~1,200 enterprise customers on the enterprise solution
- •SAP partnership: n8n available to SAP customers out-of-the-box (no separate setup/billing)
- •Enterprise examples (e.g., Mercedes) and emphasis on measurable ROI
- •AI + deterministic logic + human-in-the-loop as a value driver for enterprises
- 15:20 – 22:44
Live demo: building an email + calendar AI assistant with approvals, fallbacks, and execution trace
Jan demonstrates how n8n workflows are constructed with triggers, agents, tools, and model selection—then shows step-by-step execution logs. The demo emphasizes human approval gates for sensitive actions (e.g., sending email, creating calendar events) and model fallback to improve uptime.
- •AI assistant can generate workflows from natural language (lowering entry barriers)
- •Workflow structure: trigger node → agent → tools (Gmail/Calendar) → actions
- •Model routing: default Claude Sonnet plus fallback model for reliability
- •Human-in-the-loop approvals prevent unsafe autonomous actions
- •Audit trail: inspect inputs/outputs per node and debug by step
- 22:44 – 24:39
What reliability and auditability mean in practice (and why code-only agents can be a black box)
Aakash presses for concrete definitions of reliability and auditability. Jan explains infrastructure control via self-hosting, maintained connectors with centralized fixes, retries and deployment mechanics, and granular inspection of past executions and decision paths.
- •Reliability: self-hosting near data, fewer network dependencies, controlled runtime
- •Reliability: maintained/validated integrations—fix once, benefit everyone
- •Operational reliability: retries, stored execution data, and deployable workflows
- •Auditability: view workflow and historical executions node-by-node
- •Deterministic logic nodes make branches explainable and reproducible
- 24:39 – 29:38
Ad break: Bolt.new, Jira Product Discovery, and Product Faculty AI PM certification
Aakash shares sponsor segments about building pullable production code with Bolt.new, managing discovery-to-delivery with Jira Product Discovery, and upskilling via Product Faculty’s AI PM certification.
- •Bolt.new: generate production-grade code that engineers can extend (prototype becomes product)
- •Jira Product Discovery: discovery/prioritization/roadmapping tightly integrated with Jira
- •Product Faculty AI PM certification: agents, evals, routing, RAG, fine-tuning, guardrails
- •Discount code and course positioning for AI PM career growth
- 29:38 – 30:56
Collaboration and handoff: sharing workflows, version history, exports, and review gates
Jan explains how teams operationalize workflows: sharing access, tracking changes, and reviewing modifications before publishing. The focus is on making automation a team asset rather than an individual’s fragile codebase.
- •Share workflows with teammates via access/invites
- •Version history and publishable versions (Git-like concepts)
- •Export workflows for distribution (e.g., via email)
- •Review/approval steps for changes before publishing (governed collaboration)
- •Makes it easier to maintain and iterate on critical workflows across teams
- 30:56 – 32:40
Where n8n shines in enterprises: security orchestration, onboarding/offboarding, DevOps, and “must-not-fail” automations
Aakash challenges what’s truly unique beyond a productivity assistant. Jan outlines enterprise scenarios where failures are costly or risky—security workflows, sensitive data handling, and operational processes that demand completeness and traceability.
- •Security orchestration examples: scan attachments, route alerts, enforce handling rules
- •Sensitive processes: employee onboarding/offboarding must be correct and complete
- •DevOps/IT operations workflows with monitoring and rapid response
- •Best fit: non-personal, business-critical workflows requiring 100% certainty
- •Human-in-the-loop + deterministic steps reduce risk while still leveraging AI
- 32:40 – 39:43
What PMs should build next: start small, target repetitive cross-app work, and leverage the template library
Jan advises PMs to focus on high-frequency pain points rather than ambitious “transform the company” projects. He recommends starting with small, impactful workflows and using templates (10k+) for inspiration and quick wins that spread internally.
- •Pick tasks you do daily involving copy-paste across apps and repeated steps
- •Avoid over-scoping early; smaller workflows can deliver outsized impact
- •Not everything needs AI—deterministic automation can save major time (e.g., password resets)
- •Templates: browse by tool or function (sales/marketing/IT ops) to accelerate builds
- •Small wins drive internal evangelism and new use-case discovery across teams
- 39:43 – 41:40
Zapier vs n8n: power, flexibility, self-hosting, and deeper AI-agent primitives
Jan compares positioning: Zapier for many straightforward automations, n8n for extensibility and complex, AI-infused workflows. He highlights code nodes, richer agent configurations, memory/parsers, model fallbacks, and self-hosting as differentiators.
- •n8n emphasizes power/flexibility from the start; scale complexity over time
- •Advanced building blocks: code nodes, memory, output parsers, richer agent composition
- •Model choice and fallbacks for resilience and cost/performance control
- •Human approval gates and deterministic logic for governance
- •Self-hosting as a major differentiator versus SaaS-only approaches
- 41:40 – 44:21
Why “sprinkling AI on top” only yields 10–30%: making AI core to the value chain (and 80% AI-agent workflows)
Aakash asks about n8n’s strategic choice to make AI foundational rather than a feature. Jan explains that true leverage comes from enabling users to build agents and automations end-to-end, and notes the rapid adoption of AI agents across workflows.
- •‘Sprinkling AI’ = superficial AI buttons; limited growth impact
- •Making AI core = n8n becomes the place people build agents and automation systems
- •n8n grew ~10x by embedding AI into the product’s main value creation loop
- •~80% of workflows now use AI agents (adoption by existing tinkerer base)
- •Key pattern: AI + deterministic logic + human-in-the-loop for real production value
- 44:21 – 53:07
Business model and operating philosophy: killing lead-gen targets, avoiding per-seat pricing, and optimizing for adoption
Jan explains why n8n deprioritizes common SaaS growth levers. Profitability enables long-term thinking: maximizing usage and customer value over near-term ARR optimization, and focusing on efficient scaling rather than headcount growth.
- •Avoids per-seat pricing and lead-gen quotas to stay aligned with user value
- •Profitability reduces pressure to chase short-term ARR optics for fundraising
- •Goal shifts toward broad adoption rather than monetizing every free user immediately
- •Internal scaling goal: massive user reach with relatively small headcount
- •Treats free usage as future leverage, not a conversion funnel to squeeze immediately
- 53:07 – 1:09:03
How n8n builds product and hires AI PMs: squads, technical depth, evals, and interview signals
The final segment covers org structure and hiring: small squads, CEO involvement in key design direction, and a strong preference for technical, builder-minded PMs. Jan discusses eval ownership (AI Trust team), how candidates get noticed, and how interviews distinguish deep understanding from buzzword repetition.
- •Product org: small squads (3–5 engineers), 1–2 PMs, and a designer; lightweight overhead
- •CEO stays close via recurring design reviews while product leadership owns roadmap
- •PM bar: technical depth, system thinking, scalability/reliability trade-offs, builder mindset
- •Evals: owned by an AI Trust team; critical for model switching and consistent performance
- •Hiring: look for evidence of building + depth; prefer live problem-solving over take-homes
