I have always started things. A jazz-funk band that grew into an 11-piece and played North Sea Jazz. In 2019 two start-ups: Documole, applied AI that pulled data out of documents, in pilot talks with a bank, an insurer and a municipality years before LLMs made that easy; and Portmanto, subscription boxes of overstock fashion with paying customers and partner brands. Then I spent six years scaling other people's companies, as CRO at Naduvi and CCO at Quatt.
In 2026 the tools caught up with the ambition. Since then I build full-time with a custom-built AI agent setup. It takes an idea to a live product in days, and keeps improving it at night: content, SEO, outreach and customer service included. This allows me to test many ideas in the market, in multiple countries and languages, and double down on what works.
Every neighbourhood in the Netherlands scored on five pillars (safety, amenities, housing, social and income, nature and quiet), based on open data from various sources, including CBS, the police, RIVM and Kadaster. Interactive map, rankings, a comparison tool and a report per area.
Next.js · Python ETL into SQLite · MapLibre with vector tiles · Railway · self-hosted analytics
Looking for: real-estate, mortgage and housing platforms that want neighbourhood data or distribution.
Restaurants upload their menu; AI turns it into an allergen card for the 14 EU-mandated allergens and a multilingual online guest menu with QR code.
Next.js · Claude vision and text models · Stripe · Vercel
Looking for: hospitality wholesalers, POS and menu platforms, and trade associations to bring it to their restaurants.
Helps housing associations decide and justify, neighbourhood by neighbourhood, what to do with their commercial real estate. Builds on the neighbourhood data pipeline behind Checkjewijk.
Looking for: housing associations that want to pilot it.
Turns an installer's existing customer base and open data into qualified requests: a scoring model ranks who is most likely to buy, agents run personalised outreach and hand over warm leads. Built and launched within days; now being tested with installers.
Looking for: installers and wholesalers who want more qualified requests.
A living literature review: new studies on one clinical topic come in every week, are summarised in a fixed structure and land in a searchable knowledge base the researcher can question in plain language.
Next.js · Postgres · headless Claude worker with model routing per task
Looking for: research groups that spend too much time keeping up with the literature.
Brings decades-old VHS and Super 8 footage back to life for a private online archive: denoising, dust removal with optical flow, per-scene stabilisation and super-resolution upscaling, all running locally on Apple Silicon.
ffmpeg · OpenCV optical flow · Real-ESRGAN on Metal · GPU jobs scheduled at night
Generates the supervision roster for a department of about fifteen doctors from shifts, locations, preferences and continuity rules, with requests and changes through a chat bot.
Next.js · Postgres · one-time-password login · Telegram bot
A neighbourhood report per address for estate agents: public data on buildings, energy labels, crime, amenities and charging points, combined into a PDF with AI-written summaries and sold as a subscription. Parked to focus on Checkjewijk, which builds on the same data.
Python (Flask) · PDF generation · LLM summaries · Stripe · Railway
A childcare map with what competitors do not show: the hourly rate per location. Agents extracted 5,000+ rates from provider websites, each with source and date, across 309 landing pages in Dutch and English.
Astro (fully static) · MapLibre · agent-based web extraction
Everything above is built and run with the same setup, which I designed and use every day.
Bring the domain knowledge and the network. I build a working first version in days, and we decide together whether it is worth a company.
Founders who move fast and want a commercial co-founder who builds with AI himself, especially in consumer and energy.
Fractional or interim commercial leadership, and putting AI agents to work in marketing, sales and operations.
Want to build something together? Email me at laurenspriem@gmail.com or find me on LinkedIn.