FAQ
Frequently asked questions
Everything you need to know about Helvetic AI, our methodology, and our products.
What is Helvetic AI?
Helvetic AI is an independent Swiss AI evaluation lab. Our evaluation system answers four questions about your AI: Compliant? Performant? Reliable? Secure? Every evaluation delivers a HAAS Score across 8 dimensions in 4 pillars. Three service tiers scale from automated scores to evidence-based remediation prescriptions: Measurement, Measurement + Diagnostic, Measurement + Diagnostic + Remediation.
Does my data leave Switzerland?
No. You choose from 5 handoff modes: benchmark intelligence (standard, no data needed), API key, Docker on your infrastructure, dedicated hardware on-site, or anonymize-first. In no mode does your data leave Switzerland.
How much does it cost to get started?
Three service tiers per pillar: Assurance Basic from CHF 5,000, Assurance Plus from CHF 12,000, Assurance Komplett from CHF 20,000. An AI Risk Classification is available from CHF 3,000 as a standalone entry point. See all products on our Services page.
How long does an evaluation take?
Assurance Basic (e.g. EU AI Act Quick Check) takes about 1 week. Assurance Plus (e.g. EU AI Act Full Assessment) takes 2–3 weeks. Assurance Komplett (e.g. FINMA Alignment Check) takes 3–4 weeks. An AI Risk Classification takes about 1 week.
Do I need IT resources?
Minimal. In standard mode (benchmark intelligence), you need nothing. We already have the benchmark data. For custom evaluations, you provide an API key. The entire process is designed to minimize your effort.
What is the HAAS Score?
The Helvetic AI Assurance Score (HAAS) is our composite scoring framework across 8 dimensions across 4 pillars: Compliant (Safety, Compliance, Swiss Languages, Documentation), Performant (Performance, Robustness), Reliable (Production Reliability), and Secure (Adversarial Security). Each dimension is scored 0–100 with confidence intervals. Details on our methodology page.
What evaluation frameworks do you use?
Our evaluation system is built on three institutional frameworks: the UK AI Security Institute's evaluation framework (used by leading AI labs worldwide), the EU AI Act compliance benchmark suite developed by ETH Zurich and INSAIT, and Swiss-Bench, our proprietary benchmark suite for Swiss language and domain requirements.
What is Swiss-Bench?
Swiss-Bench is our proprietary benchmark suite with over 800 evaluation scenarios across 8 dimensions, testing models in German, French, Italian, and English on domain-specific tasks. We publish results quarterly as a public leaderboard.
What do I actually receive?
Deliverables scale with your chosen tier. Assurance Basic: automated HAAS Scores, traffic-light dashboards, and benchmark results. Assurance Plus: adds expert interpretation, gap analysis, and remediation priorities. Assurance Komplett: adds evidence-based remediation prescriptions, control mapping, and implementation guidance. Every tier includes methodology documentation for independent verification and a findings call.
How do you differ from consulting firms?
We are a technical audit lab, not a consulting firm. Our system delivers systematic, reproducible results. No subjective opinions. Entry from CHF 5,000 vs. CHF 100,000+ at Big Four. Every test is repeatable.
Are you truly independent?
Yes. No commercial relationships with any AI model provider. No referral fees, no vendor partnerships, no pay-for-score. Every model is evaluated with the same system and methodology.
What does FINMA require for AI models?
FINMA Guidance 08/2024 defines supervisory expectations for AI across 5 categories: governance, operational risk, outsourcing, data quality, and explainability. Our FINMA Alignment Check evaluates against all categories with comprehensive FINMA-mapped evaluation scenarios.
What are AI hallucinations?
AI hallucinations occur when a model generates plausible-sounding but factually incorrect information: fabricated court rulings, non-existent regulations, wrong financial data. Magesh et al. (Stanford, 2024) found that leading legal AI tools hallucinate in over 17% of queries. We quantitatively measure hallucination rates as part of the HAAS Score.
Who is behind Helvetic AI?
Helvetic AI was founded by Fatih Uenal, PhD, with the mission of making independent AI evaluation accessible to Swiss enterprises. Background: PhD (HU Berlin), Postdoc Harvard & Cambridge, MSc Computer Science (CU Boulder), MITx Statistics & Data Science. Based in Bern, Switzerland.
Is your methodology peer-reviewed?
Our methodology is grounded in 100+ peer-reviewed publications from venues including Nature, NeurIPS, ICLR, ICML, ACL, and NAACL. Our Swiss-Bench methodology is documented in two published research papers: Uenal, 2026a and Uenal, 2026b.