AI Transparency & System Card

Published in accordance with EU AI Act Article 52 (Limited Risk AI System)

EU AI ACT COMPLIANT — LIMITED RISK
Back to app

1. What ÀSÌKÒ Is

ÀSÌKÒ is a decision-support AI that helps business owners and operators analyze decisions using live macroeconomic signals, AI expert modeling, and quantitative optimization. It is not a financial advisor, legal advisor, or investment service.

Under the EU AI Act, ÀSÌKÒ is classified as a Limited Risk AI System (Article 52) LIMITED RISK. This classification applies to AI systems that interact with natural persons in a way that generates analysis, recommendations, or content — but where the human retains full decision-making authority.

Every verdict, score, and recommendation produced by ÀSÌKÒ is a decision-support output only. You are always the final decision-maker.

2. How Verdicts Are Computed

ÀSÌKÒ produces verdicts through a multi-layer analytical pipeline:

Layer 1 — Parliament of AI Experts
A panel of specialized AI agents (Risk Analyst, Market Strategist, Financial Advisor, Operations Expert, and others) each independently evaluates your decision query. Each expert votes with a recommendation and confidence score.
Layer 2 — 16 Live Macro Signals
Real-time macroeconomic indicators are ingested and weighted: Federal Funds Rate, VIX Volatility Index, Consumer Confidence, 10-Year Treasury Yield, S&P 500 trend, Unemployment Rate, PMI, CPI, GDP growth, and others. These calibrate all verdicts to current market conditions.
Layer 3 — QUBO Quantum-Inspired Optimizer
Expert votes and macro signals are combined using a Quadratic Unconstrained Binary Optimization model that finds the globally optimal verdict weighting. A classical fallback is used when quantum hardware is unavailable.
Layer 4 — Composite Score & Verdict
A 0–100 readiness score and a four-way verdict (ACT NOW / PREPARE / WAIT / AVOID) are produced. The verdict reflects the aggregated signal from all layers, not any single model's output.

3. AI Models Used

ÀSÌKÒ uses large language models (LLMs) to power the Parliament of AI Experts. The following models may be called during analysis:

GPT-4o
OpenAI
Primary reasoning engine for expert panel agents
Claude 3.5 Sonnet
Anthropic
Secondary reasoning and adversarial analysis (The Challenger)
Llama 3 70B
Meta / Groq
Fast inference for real-time macro signal processing

Model availability may vary. ÀSÌKÒ uses the best available model at the time of analysis. Your queries are never used to train any of these models.

4. Data Handling

5. Your Rights

Right to Access
Request a full export of all data we hold about you, including decision history and account metadata.
Right to Deletion
Request deletion of your account and all associated data. Processing within 30 days.
Right to Opt Out of AI
You may disable LLM-powered features at any time. Analysis will use deterministic rule-based scoring only.
Right to Explanation
You may request an explanation of any verdict. Each analysis includes a breakdown of which signals drove the outcome.
Right to Human Review
You may request that a human reviewer assess any decision output. Contact us at the address below.
GDPR Article 13
EU residents have full GDPR rights including portability and restriction of processing. See our Privacy Policy for full details.

To exercise any of these rights, contact us at jeremy@vyudu.com. We will respond within 30 days.

6. Contact

Data Controller: Vyudu Inc.
DPO / Privacy Contact: jeremy@vyudu.com
EU AI Act queries: jeremy@vyudu.com

For urgent data protection issues, please use the subject line: GDPR / EU AI ACT REQUEST. Responses within 30 days as required by law.

Last updated: June 2026 · ÀSÌKÒ v2.0 · EU AI Act Article 52 (Limited Risk) · Privacy Policy