How UDGOK Uses
Artificial Intelligence

Six Commitments That Guide Our AI Use

Transparency First

We disclose every AI tool we use, how it influences decisions, and what data it touches. No black boxes.

Human Authority

AI assists — humans decide. Every AI output is reviewed by licensed professionals before any project action is taken.

Data Protection

Client data never trains third-party AI models. We use enterprise-grade, zero-data-retention AI services only.

Fairness & Non-Discrimination

We audit our AI systems for bias in estimating, hiring, and subcontractor selection to ensure equitable outcomes.

Accountability

Every AI-assisted decision is logged with an audit trail — who used it, what was generated, what was approved.

Continuous Evaluation

We review and update our AI policies quarterly, benchmarking against NIST AI RMF 1.0 and EU AI Act standards.

How We Protect Your Data

No Training on Client Data

Your project data, drawings, and specifications are never used to train AI models — ours or anyone else's. We use only enterprise AI services with zero-data-retention agreements.

Enterprise-Grade Security

All AI interactions occur within SOC 2 Type II compliant platforms. Data is encrypted in transit (TLS 1.3) and at rest (AES-256). We never use consumer-grade AI tools on project data.

Access Controls

AI tool access is role-based and logged. Only authorized team members can interact with AI systems, and all queries involving project data are recorded in our audit log.

Retention & Deletion

AI conversation logs involving project data are automatically purged after 30 days. We maintain zero persistent memory across AI sessions — every session starts fresh with no access to prior conversations.

Vendor Agreements

Every AI vendor we use has signed a Data Processing Agreement (DPA) with explicit zero-training clauses. We audit vendor compliance annually and maintain a current vendor registry.

Regulatory Compliance

Our AI program aligns with the NIST AI Risk Management Framework 1.0, EU AI Act Article 52 transparency requirements, and CCPA data protection standards.

The Human Override Guarantee

AI will never make a final decision on your project. Every estimate, schedule, design recommendation, and quality determination passes through a licensed human professional before reaching you. If you ever have a question about whether AI influenced a deliverable, ask us — we'll provide a complete audit trail.

Policy Update Log