Selected projects.
What clients got from working with us.
Data collection
We build the pipelines that collect your market data, keep them running, and hand them over so your team runs them in-house.
Start a similar project| Competitor 01 | 1,240 € | 24 months |
| Competitor 02 | 1,310 € | 36 months |
| Competitor 03 | no offer | |
| Competitor 04 | 1,480 € | 12 months |
| Competitor 05 | 1,190 € | 24 months |
| Competitor 06 | 1,260 € | 24 months |
Financial services firm · Finance · 2025
Product positioning based on real competitor offers
A financial services firm introduced a new product and needed to know how 20 competitors make their offers: who gets one and on what terms. We built a pipeline that requests offers for 100 customer profiles and records every answer. The firm now positions its own offers against every competitor's.
20
competitors
100
profiles
| Retailer A | Retailer B | Retailer C | |
|---|---|---|---|
| Product 01 | 49 € | 45 € | 52 € |
| Product 02 | 129 € | 139 € | 125 € |
| Product 03 | 19 € | 19 € | 24 € |
| Product 04 | 89 € | 79 € | 85 € |
| Product 05 | 249 € | 259 € | 239 € |
Manufacturer · E-commerce · 2025
Pricing decisions based on up-to-date competitor pricing
A manufacturer needed the current prices of competing products across the 15 online retailers that sell them. We built a crawler per retailer and a pipeline that puts every price into one table. The manufacturer now sets its prices with every competitor price in front of it.
15
retailers
1,000
products
ML and LLM pipelines
We automate the work that slows your team down, benchmark it against real cases, and hand it over so your team runs it in-house.
Start a similar projectRear-end collision. Bumper cracked, tailgate dented, tail light broken.
1,840 €
Automated estimate
1,790 €
Expert report
Claims service · Insurance · 2025–2026
Automated repair estimates, benchmarked against real-world cases
A claims service wanted to give drivers a realistic repair estimate within seconds of reporting their damage, instead of after an expert's visit. We built a machine learning pipeline, trained on past expert cases, that takes the customer's photos and description and returns the estimate. Benchmarked against 240 expert reports, the estimates land within 10 % and 500 € of the expert's figure. Drivers get their assessment right away, and significantly more of them become customers.
<10 %
deviation
<500 €
error
Hi, when will I get the money for my claim? It was approved last week.
Answered automatically
Hello, your claim was approved on Tuesday and the payout was released today. It should arrive within two working days.
Claims service · Insurance · 2025
Repeated customer questions answered automatically
A claims service's support team lost hours every day to the same customer questions on WhatsApp. We built a knowledge base from 3,500 past chats and an agent on top of it that answers these questions on its own, and scored its replies against the team's own to make sure they match. Live after four weeks, it saves every agent two hours a day for the customers who need a person.
2 h
saved per agent
4
weeks
Technology leadership
We lead your technology from day one, build the first version, and set up the team that takes it over.
Start a similar projectLovable · Dev tools
Helsing · Defense
Parloa · Voice AI
n8n · Automation
Langdock · Gen AI
Qonto · Fintech
Family office · Venture capital · 2023–2024
Strategy, data stack and team for a new VC fund
One of Europe's largest family offices wanted to launch a venture fund that finds its deals in data instead of through its network, and had no strategy, no technology and no team for it, and processes built for a traditional family office. As interim CTO we laid out the strategy with management, brought the processes up to date, built the data infrastructure that discovers and ranks startups, and hired the team from partner to associates within three months. The fund entered the market with over 1,000 relevant deals on day one, and its team runs the stack with AI in its daily work.
1,000+
deals at launch
3
hires
| Deployments | fixed |
| Monitoring | fixed |
| Incident process | live |
| Test coverage | in progress |
| Documentation | planned |
| Data pipeline | planned |
Price comparison service · E-commerce · 2023
Plan, processes and team for a price comparison service
A price comparison service lost its CTO, with an offshore team running a platform that broke often. As interim CTO we assessed the status quo with the CEO, laid out the plan to fix the weak spots, and trained the 15-person team on the knowledge it lacked. During outages we handled the customer communication ourselves and set up the processes that prevent them going forward.
15
team members
Have a similar
challenge?
We build data pipelines, automate expert work, and lead technology teams. Tell us what you are working on and we will say whether we can help.