Different industries. Same kind of problem.
Real operational complexity, no-code hitting its ceiling, and a six-month dev project that makes no sense. Here is what I actually built.
Traders Blueprint
A custom WhatsApp inbox app for a fast-growing trading-education business.
The problem
All student support ran through personal WhatsApp accounts. Messages got lost, coaches couldn't share one inbox, and there was no record of who answered what. It worked at small scale and broke the moment they grew.
What I built
A custom WhatsApp inbox app on top of the WhatsApp Business API: one shared, multi-agent inbox where conversations get assigned, answered with saved replies, and kept with full history. The team installs it like an app and works from one place.
IT Installs
A field-service PWA technicians run their whole day from.
The problem
Technicians installing hardware on-site worked from spreadsheets and phone calls. No live job list, no proof of work, and constant scheduling confusion, including around daylight-saving shifts.
What I built
A mobile PWA technicians open on their phone: today's jobs, addresses, checklists, photo proof and live status updates, all synced. The office sees progress in real time instead of calling around, with scheduling that stays correct across timezones.
Q Bedrijfslocaties
An AI advice-report generator on a backend they actually own.
The problem
Producing commercial-location advice reports was a manual, repetitive writing job, sitting on top of a rented no-code CRM that couldn't scale with it.
What I built
A custom backend, their own PostgreSQL database with a FastAPI layer, plus an AI report generator: pull the data, generate a structured advice report in minutes, and keep everything in a system they own rather than rent.
Amazon-marketplace aggregator
An AI deal engine that runs the top of the sales funnel on its own.
The problem
A B2B sales team sourced acquisition targets by hand: researching companies, finding the right contact, and writing a first email one at a time. Slow, inconsistent, and impossible to scale with headcount.
What I built
A deal engine that finds and qualifies prospects, enriches them, drafts a tailored first email with GPT-4o, and pushes it into outreach, at a volume no person could match. The funnel fills itself, continuously.
Safwah
A subscription-and-access engine that runs itself.
The problem
A membership platform needed paid access that switches on and off automatically with subscription status, without a developer babysitting Stripe webhooks every time a payment failed or renewed.
What I built
A subscription-and-access engine: Stripe handles payments, n8n orchestrates the logic, and Airtable holds the source of truth. Sign-ups, renewals, failed-payment lockouts and reactivations all run hands-off.
Gemeente Medemblik
Turning years of HR mail into a searchable knowledge base.
The problem
A municipality's HR team answered the same employee questions over and over, with the knowledge buried across years of mailboxes and no single place to look it up.
What I built
A pipeline that pulls HR mail through Microsoft Graph and uses Claude to distil recurring questions and answers into clean, reusable knowledge-base entries, turning a quarter of a million messages into a structured set of answers.
Support Buddies
An operational ROI dashboard that proves the value in numbers.
The problem
A customer-support company couldn't show clients, in hard numbers, the value it delivered. The data existed but lived scattered across disconnected tools.
What I built
An operational ROI dashboard that pulls the numbers together and shows, per client, what the team handled and what it was worth, proof of value on a single screen instead of a story in a meeting.