Day with AI: Cold Email Pipeline from Scratch. CLI Tool, Daily Briefing, and 1-Click Draft Approval.

What I worked on

Tuesday. The whole day revolved around one problem: how to get cold emails from "we have leads in Notion" to "emails are going out and I know what's happening" — without manual copying.

What I did

Unified lead pipeline CLI

One terminal command that:

• Pulls current leads from Notion CRM
• Shows pipeline status (who's waiting, who's in follow-up, who replied)
• Connects with Clockify data — how much time I spent on each lead

Smart daily briefing

A morning message that aggregates:

• Pipeline stats from Notion
• Tracked time from Clockify for the previous day
• Today's priorities

Cold email draft preview system

This was the most interesting part. Instead of approving emails somewhere in a dashboard, I get them directly in my inbox as a preview. Each draft has three buttons:

• Approve and send (1 click)
• Edit in Notion (opens the correct record directly)
• Reject

Fix: auto-creating Clockify projects

When a lead doesn't have a Clockify project, the system now creates it automatically. Before, it crashed.

Issues along the way

Cold emails turned out to be more complex than expected. I found that:

• Personalization based on a website summary isn't enough — you need the full scrape, not a summary
• Leads belong in Lead Gen DB, not Clients DB (I had this wrong)
• I had to understand the full pipeline flow before starting implementation — otherwise I'd rewrite everything twice

Time: AI vs without AI

Task	Without AI	With AI

Lead pipeline CLI	1-2 days	~3h
Daily briefing	4h	~1h
Email preview system	1 day	~2.5h
Clockify fix + integration	4h	~1.5h
Total	2-3 days	~8h

What I learned

• Cold emails require manual personalization — scraping the full website, not just a summary. Otherwise the reader will notice
• Leads go in Lead Gen DB, not Clients DB — the difference is where they are in the process
• Before implementing, understand the full flow. Starting to code without that means rewriting