Interactive Demo

Run a live compliance risk scan. Adjust model parameters and see instant results — the same engine powering our enterprise platform.

⚡ AI Model Risk Scan

✅ What This Demo Shows

  • Model risk categorization
  • Compliance score calculation
  • Risk-level mitigation recommendations
  • Multi-factor scoring (type, scope, data, users)
  • Immediate regulatory framework matching
  • Sample mitigation roadmap preview

🚀 Full Platform Adds

  • Continuous real-time monitoring
  • Automated bias & fairness audits
  • Regulatory report generation
  • Vendor model assessment pipeline
  • Policy automation & documentation
  • Multi-model portfolio dashboards
  • Team collaboration & approval workflows
  • API integrations with ML pipelines

What You'll See

The demo runs our production risk scoring engine in real time. Here's what happens when you click "Run Risk Scan."

When you run the scan, AuditTrue evaluates your model across four critical risk dimensions and synthesizes them into a single compliance risk score. The score maps to one of three risk levels — Low, Moderate, or High — each with its own set of recommended controls, audit frequencies, and regulatory obligations.

🎯

Risk Score

A 0–250 composite score blending model type, deployment scope, data sensitivity, and user impact into one clear metric.

📋

Risk Level

Your model is categorized as Low, Moderate, or High risk — each with specific compliance obligations and audit frequencies.

🔧

Mitigation Plan

Actionable recommendations for audit trail coverage, assessment frequency, and regulatory frameworks that apply to you.

🗺️

Regulatory Map

See which regulations apply to your model — EU AI Act, GDPR, ISO 42001, US AI Executive Order — based on your parameters.

Step-by-Step Walkthrough

Follow these steps to get the most out of the interactive demo.

1

Select Your Model Type

Choose the AI model architecture closest to what you're deploying — LLM, computer vision, recommendation engine, predictive analytics, or generative AI. Each type carries different inherent risk profiles. For example, large language models score higher due to hallucination risks and broad output surface area, while predictive analytics models typically score lower unless they handle sensitive data.

2

Set Deployment Scope

Define where and how your model will be used. Internal-only models carry the least regulatory exposure. Customer-facing deployments trigger transparency obligations. Public-facing models face the strictest scrutiny, and models used in critical infrastructure (healthcare, energy, finance) fall under the highest regulatory tiers in most frameworks.

3

Choose Data Sensitivity Level

Indicate what kind of data your model processes. Non-personal data carries minimal compliance burden. Personal data brings GDPR obligations. Sensitive data (health records, financial data) triggers stricter impact assessment requirements. Biometric or behavioral data places your model in the highest risk category under most frameworks.

4

Set User Impact Scale

Specify how many people are affected by your model's decisions. Models impacting over a million users carry the highest responsibility — regulators increasingly consider population-level impact when assessing AI risk. Even a low-complexity model serving millions can be high-risk if decisions affect access to services, opportunities, or resources.

5

Describe Your Use Case

Add a brief description of what your model does. This helps the scoring engine contextualize the risk assessment. The full platform uses this description to match against known regulatory use case categories and precedent cases.

6

Run the Scan & Review Results

Click "Run Risk Scan" to generate your instant compliance risk score. The demo runs automatically on page load with default values, so you can see a sample result immediately. Adjust any parameter and re-run to see how different configurations change your risk profile — this is the same interactive analysis our platform provides across your entire model portfolio.

Try These Scenarios

Not sure where to start? Try one of these common real-world scenarios to see how different parameters affect the risk score.

🏦 Bank Loan Scoring

A predictive analytics model used to approve or deny loan applications, processing financial data for over 100K applicants annually.

High Risk

Try: Predictive Analytics → Customer-Facing → Sensitive (Health, Finance) → 100K–1M

🤖 Internal Support Chatbot

An LLM-powered chatbot for internal employee support, answering HR and IT questions. No personal customer data involved.

Low Risk

Try: Large Language Model → Internal Use Only → Non-Personal → <1,000

🏥 Medical Image Classifier

A computer vision model that assists radiologists by flagging potential abnormalities in X-ray images. Used in clinical settings.

High Risk

Try: Computer Vision → Critical Infrastructure → Sensitive (Health, Finance) → 1K–100K

🛒 E-Commerce Recommendations

A recommendation engine serving product suggestions to public consumers based on browsing and purchase history.

Moderate Risk

Try: Recommendation Engine → Public / Consumer → Personal (GDPR Scope) → 1M+

🎨 Marketing Image Generator

A generative AI tool that creates marketing visuals for internal campaigns. No user data processed.

Low Risk

Try: Generative AI → Internal Use Only → Non-Personal → <1,000

📷 Facial Recognition Access

A computer vision system using facial recognition for building access control. Processes biometric data of employees.

High Risk

Try: Computer Vision → Customer-Facing → Biometric / Behavioral → 1K–100K

Sample Output

Here's what a typical risk scan result looks like when you run the demo.

📋 Example: Customer Support Chatbot

Parameters: LLM · Customer-Facing · Personal (GDPR) · 1K–100K users

━━━ AuditTrue Risk Scan Report ━━━ Model Type: Large Language Model Deployment Scope: Customer-Facing Data Sensitivity: Personal (GDPR Scope) User Impact: 1K – 100K Composite Score: 125/250 Risk Level: MODERATE ━━━ Risk Breakdown ━━━ Model Type Factor: 35/70 Scope Factor: 30/70 Data Factor: 30/75 User Impact Factor: 20/60 ───────────────────────────── Total: 125/250 ━━━ Recommended Controls ━━━ Audit Trail: Full (automated logging) Risk Assessment: Monthly Regulatory Mapping: EU AI Act + GDPR ━━━ Mitigation Roadmap ━━━ 1. Implement automated decision logging 2. Conduct monthly bias & fairness audits 3. Generate GDPR Art. 22 assessment 4. Map to EU AI Act transparency obligations 5. Establish human oversight procedures Status: ✅ Assessment Complete

This is a simplified preview. The full AuditTrue platform generates detailed compliance reports with regulation-by-regulation mapping, gap analysis, prioritized remediation tasks, and exportable documentation for auditors and regulators.

Ready for the Full Experience?

The interactive demo shows a single-model snapshot. The complete platform monitors your entire AI portfolio in real time, with team collaboration, automated reporting, and full regulatory documentation.

Book a Full Demo →  Explore Services  See Pricing