Services

Certainty from independent Swiss infrastructure.

We help you answer the four most important questions about AI: Is our AI compliant? Is our AI performant? Is our AI reliable? Is our AI secure?

Three service pathways scale with your needs:

  • Measurement (Assurance Basic): Automated scores and traffic-light dashboards.
  • Measurement + Diagnostic (Assurance Plus): Expert interpretation, gap analysis, and remediation priorities.
  • Measurement + Diagnostic + Remediation (Assurance Komplett): Evidence-based, best-practice-proven remediation prescriptions. Control mapping, owner assignment, and implementation guidance.

We measure, diagnose, and recommend. We do not implement, validate as second line, or certify.

Jump to the right question:

Regulation & Compliance Compliant? → Model Evaluation Performant? → Production Reliability Reliable? → AI Security Secure? →

Not sure where to start? Take our Readiness Check. We’ll show you your weakest area.

Compliant?

Regulatory AI Compliance

For Chief Risk Officers, Chief Compliance Officers, and compliance teams.

Assurance Basic
EU AI Act Quick Check
Automated compliance scan against 6 EU AI Act principles. Traffic-light dashboard, key gaps identified, risk classification, Swiss regulatory overlay (nFADP + FINMA). Your starting point for regulatory clarity.
from CHF 5,000 1 week
recommended
Assurance Plus
EU AI Act Full Assessment
29 Compl-AI benchmarks × 6 principles × 18 technical requirements. Article-specific scores (Art. 9–15), full gap analysis with remediation prioritization. Incl. frontier model baselines for comparison.
from CHF 12,000 2–3 weeks
Assurance Komplett
FINMA Alignment Check
Assurance Plus + 30 FINMA-specific scenarios (Guidance 08/2024). 5 categories: governance, operational risk, outsourcing, data quality, explainability. Board-ready report with risk heat map and evidence-based, best-practice-proven remediation prescriptions.
from CHF 25,000 3–4 weeks
Performant?

AI Model Evaluation

For CTOs, Heads of AI, and engineering teams.

Assurance Basic
5-Model Evaluation
5-model comparison on Swiss-Bench: accuracy, Swiss language quality (DE/FR/IT), domain-specific scenarios, failure mode detection. Selection recommendation with evidence.
from CHF 5,000 1 week
recommended
Assurance Plus
Domain Evaluation
Assurance Basic + custom domain scenarios, task-specific evaluation, head-to-head comparisons with confidence intervals. System reliability metrics included.
from CHF 12,000 2–3 weeks
Assurance Komplett
Full SOTA Sweep
Coming Q4 2026
30+ models evaluated. Swiss-Bench + Compl-AI + custom domain. Full ranking table, TCO analysis, Swiss language quality, compliance sidebar, and evidence-based remediation prescriptions. The definitive comparison.
Pricing on request 3–4 weeks
Reliable?

AI Reliability Evidence

Independent evidence for three decisions: which model to pick, what to put in your governance file, and what's actually wrong when something fails in production. For CTOs, engineering teams, CCOs, and reviewing committees. Swiss financial institutions are our deepest vertical; the dedicated edition is in active development for Q3 2026.

Selection
Model Reliability Check
Coming Q3 2026
Compare candidate models head-to-head before you commit. The general edition, suitable for any industry. Same scoring as the Swiss FI Edition; the general probe library is still in development.
CHF 5,000–18,000 1–2 weeks
Selection · Swiss FI Edition
Swiss Model Reliability Check
Coming Q3 2026
Compare candidate models on what actually matters in a Swiss bank: CHF/EUR precision, capital ratios, AML escalation, FINMA terminology. 96 probes across DE / FR / EN, scored deterministically.
CHF 5,000–18,000 1–2 weeks
Assessment
AI Reliability Evidence Pack
Coming Q3 2026
A document set you can attach to your governance file before approval, audit, or supervisory review. Evidence ledger, 12-item governance check, independent assessment memo with references to FINMA Guidance 08/2024 and related rules where applicable, reproducibility packet. For CCOs, Heads of Risk, and Internal Audit.
CHF 8,000–28,000 2–4 weeks
Diagnose
RAG Production Diagnostic
Coming Q3 2026
Find out why your live RAG or agent system is misbehaving. Failure classified against ten known modes; replay of the exact text the model saw; concrete remediation steps with acceptance criteria. Two reports from one engagement: engineering view and a one-page release-readiness view for the CCO.
CHF 12,000–28,000 3–5 weeks
Secure?

AI Security & Adversarial Testing

For CISOs, security teams, and SOC analysts.

Assurance Basic
Automated Security Scan
Prompt injection testing, data leakage detection, jailbreak resistance. Automated test suite, report with pass/fail per attack vector.
from CHF 5,000 1 week
recommended
Assurance Plus
AI Penetration Testing
Assurance Basic + adversarial robustness testing, multi-step attack chains, MITRE ATT&CK mapping for AI. Detailed vulnerability report with remediation prioritization.
from CHF 12,000 2–3 weeks
Assurance Komplett
Full Red Team Exercise
Coming Q4 2026
Assurance Plus + CALDERA-based adversary emulation, autonomous attack agents, detection validation, purple team assessment. Full MITRE ATT&CK report with evidence-based remediation prescriptions.
Pricing on request 3–4 weeks
Cross-Cutting

Not sure where to start? We have three options.

Entry
AI Risk Check
Risk classification (EU AI Act + FINMA + nFADP). Determines which of the four questions is most urgent for your AI system and recommends the right evaluation path.
from CHF 3,000 1 week
Integrated
Integrated AI Assurance
Coming Q4 2026
All four questions at Assurance Plus tier in one engagement. One report covering compliance, performance, reliability, and security. The most comprehensive evaluation on the Swiss market.
Pricing on request 6–8 weeks
Recurring
AI Monitoring
Coming 2027
Quarterly re-evaluation across all four dimensions. Drift detection, compliance tracking, reliability regression, security posture.
Pricing on request Quarterly

Or take our Readiness Check. We’ll show you your weakest area.

On Request

Additional services on request

Implementation
AI Model Router
Coming Q4 2026
Your evaluation results as executable routing rules. Three variants: Config (JSON/YAML), SDK (Python with audit logging), or Proxy (Docker, OpenAI-compatible). Automatic model selection, compliance guardrails, cost optimization.
Pricing on request 1–3 weeks
Advisory
Local AI Setup Advisor
Coming Q4 2026
Run AI models locally instead of in the cloud? Hardware and software recommendation, model selection per use case, 3-year TCO comparison (local vs. cloud), security checklist for on-premise AI.
Pricing on request 1–2 weeks
Specialized
Helvetic AI Select – Verified Library
Pre-verified domain models tested on Swiss-Bench. Benchmark report with base model comparison, Swiss language evaluation, EU AI Act compliance assessment, deployment guide. Cybersecurity, finance, and medical domains available.
from CHF 5,000 1–2 weeks
Specialized
Helvetic AI Select – Custom Fine-Tuning
Coming Q4 2026
Your domain dataset, our sovereign hardware. LoRA/DoRA fine-tuning on Swiss infrastructure, delivered with full benchmark comparison against base model. Zero data leaves Switzerland.
Pricing on request 2–4 weeks
Resources

Free. Because transparency builds trust.

Quarterly Benchmark
Swiss-Bench Leaderboard
800+ proprietary evaluation scenarios. DE/FR/IT. How do frontier models perform on Swiss-specific tasks? Updated quarterly.
View results →
Instant Result
AI Readiness Check
Short questionnaire across all four dimensions. Find out where your greatest need for action lies, with a concrete product recommendation.
Start check →
Free Download
Quarterly Report
Swiss-Bench + compliance suite against 10–15 frontier models. Compliance scores, Swiss language performance, FINMA alignment.
Request report →
These resources demonstrate our methodology. The same system produces your individual evaluation.
Data Handoff

5 delivery tiers. You choose.

Tier Mode How It Works For Whom
Intelligence Benchmark Intelligence DEFAULT We benchmark commercial models with our own keys. You receive scores + interpretation. No access needed. Default for everyone
Standard API Key You provide endpoint for custom/self-hosted models. System runs on our infrastructure. Custom model evaluation
Regulated Docker Dockerized system runs on your infrastructure. You send output logs. Banks, insurers (FINMA)
Premium On-Premise We deploy hardware on your site. Full air-gap. High-security, classified
Privacy-First Anonymize-First You run anonymization script (we provide). Send anonymized data. We evaluate. Data-sensitive clients
Sovereign AI Lab. Open-source and open-weight models run on our own hardware in Switzerland at reference and production quality. Proprietary models are evaluated via their providers’ APIs. Your data never leaves Switzerland. Methodology →
Contact

Ready for an independent evaluation?

Start with an AI Risk Check (from CHF 3,000) or go straight to the question that concerns you most. Within one to two weeks you’ll know where your AI systems stand. Evidence-based, not opinion-based.

Assurance Basic from CHF 5,000 · Assurance Plus from CHF 12,000 · Assurance Komplett from CHF 20,000