Enock

Day 3 of building in public 🧵 I built the same automation TWO different ways today. Way 1: One Claude prompt per platform → direct, beginner-friendly Way 2: One master prompt → JSON parse → Router → cleaner, but harder Same result. Very different complexity. #BuildInPublic

Want to swap Google Docs for Notion? Ask the Maia agent how the scenario is structured. It'll explain the connections so you can adapt it. Building in public continues. #MakeAutomation #AIAgents #BuildingInPublic

This replaces: ❌ Opening multiple docs manually ❌ Copy-pasting between files ❌ Remembering which doc has what With: ✅ Conversational commands ✅ Instant document operations ✅ Single interface for everything

Tweet 1: Day 14: Built a Document Agent in Make.com One chat interface. Three actions: Create documents from scratch Read existing files Update content on command No more switching between Google Docs tabs.

The solution: Make AI Agents + Gmail + Google Calendar + Claude Watch inbox → Auto-research → Qualify lead (HIGH/MEDIUM/LOW) → Propose meeting times → Draft reply → Notify Slack Fully autonomous. Zero human input. #AI #AIAutomation #Make #Claude #BuildInPublic

The problem: Leads hit my inbox. I manually research their company, check my calendar, write a custom reply. 45 mins per lead. That's unsustainable when you're trying to scale consulting.

Day 13: Built a Sales Outreach Agent that killed my biggest bottleneck Gmail → Research company → Score lead → Check calendar → Draft personalized reply The trick? Cut the agent prompt from 1,400 words to 400. Less instructions = more reliability. Full breakdown 🧵

The biggest insight: Agents are decision-makers. This agent chooses the structure before it writes the first word. Yesterday was about research; today is about execution. Day 13 is up next. Save this if you're building agentic workflows! 🔖

What I solved: ✅ Blank page syndrome (Instant drafts) ✅ Manual filing (Auto-logs to Sheets/Docs) ✅ Context switching (Everything stays in Slack) ✅ Formatting hell (Ready-to-edit Doc links)

Day 12 of building in public ⚡ I just leveled up the AI stack. One Slack trigger ➔ 4 autonomous tools. Meet the "Content Draft Creator"—my new AI teammate that researches, writes, and files its own paperwork. 🧵👇 #BuildingInPublic #AIAgents

The biggest insight from building this: With an Agent, you just give it an objective, hand it a toolkit, and let it figure out the execution. One form. Four tools. Zero manual research. Save this if you're building AI agents too 🔖 #BuildingInPublic #AIAgents #MakeAutomation

What I solved with this system: ✅ The manual research problem — hours of work condensed to a 30s form ✅ The rigid automation problem — I don't have to map every single 'if/then' step ✅ The messy data problem — unstructured web data becomes clean database rows

Day 11 of building in public ⚡ Today I graduated from basic automations to my first autonomous AI Agent. One trigger. Four tools. Fully autonomous. Here's the full system 🧵👇

Day 10 of #BuildingInPublic: I just retired my manual crypto price checking. ☕️📈 I built an automated "Morning Brief" that hits my Telegram at 7 AM daily. The Stack: ⏰ Schedule (Make.com) 🌐 HTTP (CoinGecko API) 🟪 Claude (AI Analysis) ✈️ Telegram (The UI) #AI

Day 9 of building in public 📰 Built a 2-pipeline newsletter machine today. Pipeline 1 → Every Monday: RSS feeds + Claude writes a research newsletter Pipeline 2 → Every Friday: My social posts compiled into a Builder's Log digest Two newsletters. Two audiences. Zero blank