The speed of development is accelerating faster than most people realize. In a single week, I shipped (or helped ship) five separate projects using AI tools — and each one would have taken significantly longer without them.
Here’s the breakdown of what I built, the tools I used, and the real time savings:
The Five Projects
1. Custom AWS Lambda Layer — OpenAI Codex CLI (saved 4+ hours)
I needed a Lambda Layer with specific Python dependencies for a data pipeline. Instead of manually wrestling with Docker, pip, and zip packaging, I described what I needed to Codex CLI and iterated on the output. The layer was built, tested, and deployed in under an hour. Normally this is a half-day task just fighting dependency compatibility.
2. Looker LookML Debugging Query Suite — Cursor Pro (saved 2+ days)
LookML debugging is notoriously tedious. I used Cursor Pro to generate a comprehensive suite of validation queries that check for common LookML issues — broken joins, circular dependencies, unused views, and type mismatches. What would have been two days of writing and testing SQL became an afternoon of iterating with AI suggestions.
3. Voice-to-Text Enhancement Workflow — Voice Ink + Groq (saving 30+ min/day)
I set up a workflow where Voice Ink captures spoken notes and Groq cleans them up into structured text. This isn’t just transcription — the AI reformats rambling voice notes into clear bullet points, extracts action items, and fixes the verbal tics that creep into dictation. The time savings compound daily.
4. Google Sheets Automations for Non-Technical Coworkers — ChatGPT (saving hours/month)
Two coworkers who don’t write code needed to automate repetitive spreadsheet tasks. I sat with each of them, taught them how to describe what they wanted to ChatGPT, and helped them iterate on the Google Apps Script output. Both now have working automations they maintain themselves. The multiplier effect here is the real story — I spent an hour, they each save hours every month.
5. Chrome Keystroke Automation for My Husband — ChatGPT + Tampermonkey (saving 1+ hr/day)
My husband had a repetitive browser-based workflow that involved the same clicks and keystrokes hundreds of times a day. I taught him how to prompt ChatGPT to build a Tampermonkey script that automates the sequence. He’s not a coder. He now saves at least an hour daily. He maintains and updates the script himself by chatting with AI.
The Takeaway
The common thread across all five projects isn’t the specific tools. It’s the compound effect of looking for automation opportunities everywhere, every day.
The gap between people who actively integrate AI into their workflows and those who don’t is widening fast. It’s not about replacing your job — it’s about doing your job at a fundamentally different speed. The people I helped aren’t technical. They just needed someone to show them the door.
If you’re not looking for ways to improve your workflow with AI daily, you are falling behind. Not theoretically. Measurably.