From "How Are My Ads Doing?" to Complete Reporting Infrastructure
I asked for campaign stats. My AI built an entire reporting system I can query anytime.
I run Google Ads for OsteosOnline—my network of osteopathy practices across Belgium and the Netherlands. Hundreds of campaigns. Thousands in monthly spend. I wanted to know: How are we doing this month? Which campaigns are performing? Where's the waste? Logging into Google Ads every time? Opening dashboards? Exporting spreadsheets? No. "Build me a reporting system I can query from the terminal."
The Request
I wanted four things: 1. Campaign overview — All campaigns with clicks, conversions, cost, CPA 2. Month-over-month comparison — This month vs last month, with trend arrows 3. Top performers — Which campaigns are winning? 4. Problem detection — Which campaigns have the highest cost-per-acquisition? Not a dashboard I have to visit. Not a weekly email I'll ignore. A tool I can query whenever I think about it.
What Got Built
My colleague built a complete reporting CLI:
- Google Ads API integration — Authenticated via OAuth, reads from my actual account (customer ID 6391578534)
- Four report commands — campaigns, compare, top, bottom. Each answers a different question
- Month comparison logic — Automatically calculates this month vs previous month, shows percentage changes
- CPA analysis — Identifies campaigns with worst cost-per-acquisition so I know where to cut
- Clean terminal output — Formatted tables, emoji indicators (📈📉), totals calculated
What It Looks Like
Now when I want to know how my ads are doing: $ python ads_report.py campaigns ✅ OsteosOnline NL - Amsterdam Clicks: 1,247 | Conversies: 89 | Kosten: €412.50 | CPA: €4.63 ✅ OsteosOnline BE - Antwerpen Clicks: 892 | Conversies: 67 | Kosten: €298.30 | CPA: €4.45 ... TOTAAL: 8,432 clicks | 612 conversies | €3,241.80 Gemiddelde CPA: €5.30 Or I run 'compare' and see month-over-month trends. Or 'top' for winners. Or 'bottom' to find problems. No dashboards. No exports. Just answers.
The Reality
Before: I'd forget to check ads for weeks. Then spend 30 minutes clicking through Google Ads trying to figure out what's happening. Now: I think about ads, I run a command, I have my answer in 3 seconds. The system compounds. Every question I ask, my colleague adds to the tool. "Can you also show me which ad groups are underperforming?" Done. Now it's part of the system. This isn't a one-time report. It's infrastructure that grows.
Why This Matters
Most people treat AI as a question-answering machine. Ask once, get answer, forget. I treat AI as an infrastructure builder. Ask once, get a system, use forever. The Google Ads reporting tool took one conversation to build. Now I have instant access to my campaign performance, anytime, from anywhere. That's the difference. Not "AI helped me once." But "AI built me something permanent." €3,000+ monthly ad spend, monitored with a single command. That's what AI infrastructure actually delivers.
Want to build something like this?
I help people design AI infrastructure that actually works. Not chatbots. Not wrappers. Real systems.
Let's talk