# Dataroots

> Dataroots is a reviewed AI agency in Ghent, Belgium, listed in the AIAgencies.eu register and last assessed 2026-09-07.

Canonical page: https://www.aiagencies.eu/agency/dataroots
Generated: 2026-09-23 · Curated by Gabor Kiss · No paid placements — inclusion, score, and rank are never sold.

## Facts
- Location: Ghent, Belgium
- Website: https://dataroots.io
- Hourly rate: not published
- Team size: 50-249
- Founded: 2016
- Services: AI Consulting, AI Development, Generative AI
- Industries: FinTech, E-commerce, SaaS & B2B
- LinkedIn: https://www.linkedin.com/company/dataroots

## About
Dataroots is a data and AI consultancy and part of the Talan Group, working across data strategy, machine learning, MLOps and generative AI for enterprise clients, with a Ghent office and its head office in Leuven.

## Documented clients
Documented engagements: KBC, AB InBev, and Toyota. All 3 carry a source URL on file.

## Curator assessment
Reviewed 2026-09-07 by Gabor Kiss against published evidence.

Dataroots publishes no client case studies—the site carries a customer logo wall (KBC, AB InBev, Telenet, Proximus, Toyota) with no engagement description behind any of it.

What it does publish is engineering writing: a weekly technical blog, a podcast, open-source contributions, and a build note on a conversational data assistant that names the whole stack—OpenMetadata, Amazon Bedrock, Cohere embeddings, AWS Athena, Airflow and a six-agent pipeline—while stating plainly that it started as an internal proof of concept.

The firm is now dataroots, a Talan company, and absorbed Micropole Belgium's teams on 1 January 2026, with 144 people published on its DNA page, so the delivery organization behind a proposal today is larger and newer than the 2016 founding date suggests.

Best for buyers who can evaluate a data and AI partner on engineering writing, team scale and a reference call rather than on published outcomes—if you need to read what was deployed for a comparable client, that evidence is not on the site.

### Key strengths
- Technical writing carries real depth—architecture, tool choices and tradeoffs are explained, including an LLMOps piece arguing that a demo is not production
- 144 people published with certification counts, and Talan Group ownership adds delivery capacity across Data, AI and SAP after the Micropole Belgium integration
- The published generative-AI build labels itself a proof of concept instead of dressing it as a client deployment

### Good to know
- Zero published case studies—ask for two written engagement summaries with outcomes and a reference client you can speak to
- The blue-chip logos carry no described work behind them—confirm what dataroots specifically delivered for the logo closest to your use case
- Ownership changed twice in a year—ask who staffs your team, from which legacy organization, and at which office

---
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