PortfolioEU-sovereign · EU-only hosting
AI Governance · EU AI Act
Compliance-by-Proof for AI systems.
Art. 13Transparency & information
PASSEDArt. 14Human oversight
PASSEDArt. 15Robustness & accuracy
TESTEDArt. 12Record-keeping obligations
EVIDENCEAuditable evidence per article — not self-attestation
01 — Why
LUMENAAItestsAIsystemstechnicallyagainstspecificEUAIActarticlesandproducesauditableevidence—compliance-by-proofinsteadofself-attestation.Anditproposes,basedondata,whichAIiseconomicallyworthwhile.
02 — Problem space
Thegovernancegap
Scan 01
Self-attestation is not evidence
Questionnaires and policies do not demonstrate how an AI system actually behaves. The EU AI Act demands more than good intentions.
Scan 02
Articles are specific — tests are rare
Transparency obligations, robustness, human oversight: the requirements are precisely worded, yet rarely tested technically.
Scan 03
Which AI is even worth it?
Before the governance question comes the investment question. Without a data basis, use-case selection remains gut feeling.
03 — How it works
Twowedges,onesystem
Compliance-by-proof and use-case ROI interlock: first the technical evidence, then the economic decision — and LUMENA governs the in-between.
01
Connect
The AI system is connected technically — EU-sovereign, EU-only hosting.
02
Test
Specific EU AI Act articles are tested technically, not surveyed.
03
Generate evidence
Every test produces auditable, traceable evidence.
04
Evaluate use cases
Data-based proposals for which AI pays off, at what ROI.
05
Govern
LUMENA governs the in-between — from idea to compliant operation.
04 — Principles
01Compliance-by-proof: technical tests against specific articles, not checklists.
02EU-sovereign — EU-only hosting, operator-blind architecture.
03Auditable evidence as the result, not a by-product.
04Wedge 2: use-case generation with ROI — which AI is truly worth it?