Writing

Notes & Thoughts.

Essays on AI automation, agents and enterprise AI: what holds up in production, what quietly breaks, and how to build systems a business can depend on. Each piece first ran on the TrueHorizon AI blog.

Latest essay

All essays

17 essays · 75 minutes of reading · newest first

  1. Claude Design for Enterprises: 6 Things to Know Before You Prototype

    It is not a prettier Canva. Design system extraction, the four-part prompt framework, and an honest map of where it still breaks.

    Enterprise AI6 min readFirst published on TrueHorizon AI

  2. The AI Agent That Does YouTube Research While You Sleep

    Viral outlier detection, transcript analysis, and thumbnail vision models: a content research pipeline built on validated data instead of guesswork.

    AI Agents5 min readFirst published on TrueHorizon AI

  3. Stop Prototyping Slowly: n8n MCP + Claude Code

    From plain-English prompt to production-shaped workflow in about 20 minutes. Why MCP changes what Claude actually knows about your stack, and where humans still matter.

    Automation5 min readFirst published on TrueHorizon AI

  4. Why AI Chatbots Fail: It's a Context Problem, Not an Intelligence Problem

    Most AI support failures are blindness, not stupidity. The context-first architecture I build instead, with human approval kept in the loop.

    AI Systems6 min readFirst published on TrueHorizon AI

  5. AI Makes You Faster. Does It Make You Worse?

    Speed is measurable and mastery is not. On using AI for leverage without quietly outsourcing the judgment that made you good.

    Perspective3 min readFirst published on TrueHorizon AI

  6. Project Glasswing: Why AI Regulation Can't Wait Any Longer

    AI already decides credit, diagnoses, and hiring while the rules move at a different speed. What tiered, adaptive governance should look like.

    AI Governance5 min readFirst published on TrueHorizon AI

  7. Lead Qualification Is the Real Outbound Bottleneck

    Pointing AI at writing emails faster only accelerates rejection. The actual constraint is deciding who deserves the email at all.

    Revenue Automation4 min readFirst published on TrueHorizon AI

  8. No, n8n Isn't Dead. It Turned Into Infrastructure.

    The hype went quiet and people read that as decline. The tourists left and the builders stayed, which is what maturity actually looks like.

    Automation4 min readFirst published on TrueHorizon AI

  9. Why 2026 Belongs to Autonomous AI Agents

    Systems that hold a goal, plan the steps, and execute across platforms are moving from demo to deployment. What that does to audits and to data entry roles.

    AI Agents4 min readFirst published on TrueHorizon AI

  10. AI for Accountants: Automating Audit and Tax in 2026

    Continuous reconciliation turns the audit into a verification layer rather than a discovery exercise. What that means for the people doing the discovering.

    Enterprise AI3 min readFirst published on TrueHorizon AI

  11. Claude Code Review and the End of the PR Bottleneck

    Specialised agents that cross-check each other, comment inline, and scale review depth to the risk of the change.

    Engineering4 min readFirst published on TrueHorizon AI

  12. Claude Code Plus n8n: How One Person Covers Four Jobs

    One tool makes you a faster builder, the other makes you present in more places at once. Together they change the arithmetic of headcount.

    Automation4 min readFirst published on TrueHorizon AI

  13. Why n8n Is Ready to Replace the Enterprise Automation Giants

    Enterprise-grade has been marketing language for expensive contracts. The honest definition is control over where workflows run and what happens to your data.

    Enterprise AI4 min readFirst published on TrueHorizon AI

  14. How n8n Multiplies Revenue in the Age of Agentic AI

    Scripted automation executes steps. Agentic systems reason and adapt. That difference shows up in velocity, in margin, and in signals you used to catch too late.

    Revenue Automation4 min readFirst published on TrueHorizon AI

  15. Moltbook's Hidden Promise: Platforms That Optimise for Understanding

    Most platforms optimise for time on site. Optimising for whether you understood something would produce a very different content ecosystem.

    Perspective4 min readFirst published on TrueHorizon AI

  16. Closing the AI Skills Gap in an AI-First Workplace

    Companies bought the infrastructure and found their teams could not use it. The gap is conceptual, and the fix is training for partnership.

    Future of Work5 min readFirst published on TrueHorizon AI

  17. Most Fortune 500s Are One Audit Away From an AI Revolution

    Enterprises bleed money through process inefficiency that never reaches an earnings call. The audit is what turns invisible cost into a roadmap.

    Enterprise AI5 min readFirst published on TrueHorizon AI

What I am writing next.

Shorter pieces from the build, in progress now.

  • Systems Why most AI automations die in production
  • Automation n8n or code? A decision framework
  • Agents Designing AI agents that actually ship
  • Document AI From PDF chaos to clean data: document intelligence
  • Strategy Scaling output without scaling headcount
  • Enterprise What Fortune 500 teams get wrong about AI

More writing is on the way. Want me to write about something specific?