AI & data
AI that earns its place in the product
From predicting partner churn to replacing averages with per-lead machine learning, and knowing when to stop a pilot. These are the moments where AI and data changed how we worked, and what they delivered.
Hands-on experiment
How I built this site
I built this site myself with Lovable, to learn what it takes to run AI in production rather than in a demo.
Every morning at 09:00 CET a scheduled job reads the latest posts from a handful of sources I chose, such as Lenny's Newsletter, Mind the Product, Roman Pichler, The Product Guy and TechCrunch. A language model is instructed to look for a pattern across at least three of them and write an original opinion piece in my voice, based on my profile, while avoiding topics from recent columns. Another model generates an abstract header image. The column is stored in the site's database and published automatically.
The cases library follows the STAR structure and is also database-backed, so I can draft, edit and publish each case separately.
What I learned: generating is easy. Quality control, avoiding repetition and keeping it sounding like me are the hard part, and still need human review.
From Averages to Per-Lead Predictions
I replaced a static return-rate average with an ML model that improved revenue forecasts, strengthened confidence in the numbers and enabled region-level bidding.
Predicting Partner Churn Before It Happened
I introduced churn and lifetime-value prediction so our partner happiness team could identify and support at-risk partners earlier.
Automating Regional Ad Targeting
I built a targeting flywheel that aligned regional ad spend with partner coverage and revenue per lead, helping paid marketing grow revenue by 28% in 2023.
Building a Team-Owned Experimentation Practice
I built a team-owned experimentation practice that achieved an 11% win rate, generated €1.2M in additional revenue and uncovered a key measurement issue.
Piloting AI With Honest Judgement
I helped mature two AI products while identifying where automation added real value—and where it did not.
Killing an Idea with Data
My analysis showed that reselling unsold leads was worth far less than assumed, so we prioritised better regional targeting instead.
Killing My Own AI Golf Coach Idea
I validated an AI golf coach before investing further and stopped when the market and cost model did not support a serious business.