# Automating the Boring Parts: How I Built an AI-Powered Blog Agent

*By [Gianfranco Mileo](https://gianfranco-mileo.com/) · Published January 17, 2026 · updated September 6, 2026 · 2 min read · Category: AI & Technology · Tags: AI Agents, Automation, Productivity · Canonical: https://gianfranco-mileo.com/blog/automating-the-boring-parts-how-i-built-an-ai-powered-blog-agent*

> I built an AI agent that writes weekly blog posts about AI in program management. Here's the architecture, the decisions, and what I learned about AI content quality.

## Why Automate Blog Writing?

As a TPM at Google, I don't have unlimited time to write blog posts. But I do want to consistently share insights about how AI is changing program management. So I did what any engineer-turned-TPM would do: I built a system to help.

This website has an AI agent that runs every Monday morning. It fetches the latest AI news, generates a draft blog post from my perspective, and emails it to me for approval. If I don't respond within 48 hours, it auto-publishes. If I reject it, it's archived and I can edit it later.

## The Architecture

The system is surprisingly simple:

1. **News Aggregation** — RSS feeds from Google AI Blog, TechCrunch AI, MIT Technology Review, and VentureBeat are parsed weekly
2. **Content Generation** — Gemini 2.0 Flash receives the news summaries plus a detailed system prompt about my voice, background, and preferred topics
3. **Human-in-the-Loop** — An email with approve/reject buttons gives me editorial control without requiring me to log into a CMS
4. **Auto-Publish Safety Net** — If I'm busy or traveling, the post still goes out. It's a feature, not a bug — consistency matters more than perfection for a blog.

Related: see [Lessons from Scaling a Company to 75 People — And What AI Would Have Changed](https://gianfranco-mileo.com/blog/lessons-from-scaling-a-company-to-75-people-and-what-ai-would-have-changed) and [The Real AI Bottleneck Isn't the Agents — It's the Wiring Between Them](https://gianfranco-mileo.com/blog/the-real-ai-bottleneck-isnt-the-agents-its-the-wiring-between-them).

## What I Learned About AI Content

The raw AI output is about 70% of the way there. It captures the facts and structures them well, but it lacks the specific anecdotes and opinions that make content personal. I typically spend 10-15 minutes editing each draft — adding a real example from my work, adjusting the tone, and cutting the generic parts.

That 15-minute investment is dramatically more efficient than writing from scratch, which would take me 2-3 hours. It's the perfect division of labor: AI handles research and structure, I handle voice and experience.

> The goal isn't AI-generated content. It's AI-amplified content — where the ideas are mine but the heavy lifting of research, drafting, and formatting is automated.

---

Original article: https://gianfranco-mileo.com/blog/automating-the-boring-parts-how-i-built-an-ai-powered-blog-agent  
More articles: https://gianfranco-mileo.com/blog · RSS: https://gianfranco-mileo.com/feed.xml · JSON Feed: https://gianfranco-mileo.com/feed.json · Site context for agents: https://gianfranco-mileo.com/llms.txt