Prove your AI training data complies with the AI Act
You hand us your raw data. You walk away with an auditable dossier proving it complies with Article 10 — provenance, quality and bias measured. Processing happens on your side: your data doesn't leave.
For organisations that build, train or fine-tune AI — not those that merely use one.
- Operated in the EU
- In-place — your data never leaves
- Manifest proven at runtime
How it works
Scoping
We classify your system's risk. Often the useful conclusion is: “you don't need all of this.”
In-place analysis
Provenance, quality, PII and bias, measured automatically. Your raw data never leaves your perimeter.
Auditable dossier
In Article 10 format, exportable for the regulator, and kept up to date as your data evolves.
The deliverable: an Article 10 compliance dossier
Datasheet, provenance, quality, bias, governance — the evidence you present during a conformity assessment.
Already have an AI governance platform? Here's the tab it leaves empty →
From €500/month (excl. VAT), no commitment.
A software product that scales — not consulting billed by the hour.
See the offerFrequently asked questions about the AI Act
What is Article 10 of the AI Act?
For high-risk AI systems, Article 10 requires governance of training, validation and test data: relevance, representativeness, traceable provenance, bias examination and documentation. It is the core of the AI Act's data-compliance requirements.
Am I concerned by the AI Act?
It depends on two things: your role and your risk level. On role, Article 10 (the training-data dossier) only applies to AI builders/providers; if you merely use a third party's AI (deployer), that obligation falls on your provider, not you. On risk, most systems fall under minimal or limited risk, with few obligations; only high-risk systems carry the bulk, including Article 10. Many organisations overestimate their obligations.
What is the AI Act timeline?
The AI Act entered into force in 2024 and its obligations apply in phases depending on the system type, with deadlines adjusted (Digital Omnibus). As the timeline evolves, see our dedicated article for up-to-date dates.
AI Act vs GDPR: what differs on training data?
GDPR governs personal data (legal basis, purpose); the AI Act governs the quality and governance of training datasets. They overlap on provenance and profiling, but pursue distinct goals and apply in parallel.
What is an Article 10-compliant data dossier?
Auditable documentation of the dataset: origin and legal basis, composition, preparation, quality, bias, gaps and governance — producible as evidence during a conformity assessment.
I already have an AI compliance / governance platform. How is Conformlex different?
Most AI Act tools are governance platforms: they document and orchestrate compliance (registry, controls, regulator-ready reports) but don't measure your data. Yet Article 10 requires evidence at the dataset level — provenance, quality, bias measured. That's what Conformlex produces, as a complement to your platform: the dossier attaches to it, it doesn't replace it.