# Datahouse

> Datahouse is a reviewed AI agency in Zurich, Switzerland, listed in the AIAgencies.eu register and last assessed 2026-08-07.

Canonical page: https://www.aiagencies.eu/agency/datahouse
Generated: 2026-09-21 · Curated by Gabor Kiss · No paid placements — inclusion, score, and rank are never sold.

## Facts
- Location: Zurich, Switzerland
- Website: https://www.datahouse.ch
- Hourly rate: not published
- Team size: 10-49
- Founded: 2005
- Services: AI Development, AI Consulting
- Industries: FinTech
- LinkedIn: https://www.linkedin.com/company/datahouse-zurich

## About
Datahouse is a Zurich data-science and software firm (an ETH Zurich spin-off) building custom machine learning and data solutions for real estate, finance and insurance clients.

## Documented clients
Documented engagements: Vaudoise, Wüest Appraisal, Globalance Bank, Zürcher Kantonalbank, and Scout24 Schweiz AG. 3 of 5 carry a source URL on file.

## Curator assessment
Reviewed 2026-08-07 by Gabor Kiss against published evidence.

Five published client stories cover Vaudoise, Globalance Bank, Wüest Appraisal and an anonymized psychiatric-assessment digitization, but only one describes AI work in any technical depth.

The strongest technical writing sits in the blog rather than the case library: a transfer-learning build on MobileNet sorting real-estate images into eleven categories at 85% accuracy, published with a public test tool—genuine method, but dated 2021 and not attached to a named client engagement.

Best for buyers in Swiss insurance, banking or real-estate analytics who want a small ETH-descended data-science team and are comfortable assessing capability from method writing rather than from outcome metrics.

### Key strengths
- Named clients include Zürcher Kantonalbank, Scout24 Schweiz and Vaudoise—substantial institutions for a team of this size
- Hedonic property valuation work for Wüest Appraisal sits in a genuinely quantitative niche
- Publishes its methods openly, including a working public demo of the image-classification model

### Good to know
- Ask what has shipped since 2021—the most technical published AI work is now several years old
- Most case studies describe data and software delivery rather than machine learning; confirm which your project would be
- A ten-to-fifty person team means key-person exposure—ask who specifically would do the modeling

---
Source: AIAgencies.eu — the curated register of European AI agencies. Rate provenance: "agency-confirmed" = disclosed via submission or claimed profile; "directory-listed" = Clutch or two agreeing B2B directories (lone unverified sources are never written). Full method: https://www.aiagencies.eu/methodology · Rate index: https://www.aiagencies.eu/rates