AI Disclaimer

Effective Date: March 22, 2026  |  Last Updated: March 22, 2026

SupplyChainStack, operated by Steeled Inc., a Delaware Corporation, uses artificial intelligence and machine learning technologies to analyze supply chain data, generate forecasts, identify risks, and provide recommendations. This AI Disclaimer explains how our AI-generated content works, its inherent limitations, and your responsibilities when using it.

What Is AI-Generated Content?

Throughout the SupplyChainStack platform, you will encounter outputs generated by artificial intelligence, including:

• Demand forecasts and sales predictions based on historical data;
• Inventory health assessments and reorder recommendations;
• Supplier risk scores and evaluation summaries;
• Cost optimization suggestions and landed cost estimates;
• Risk alerts and disruption probability assessments;
• Marketplace match scores between buyers and service providers;
• Freight rate estimates and logistics cost projections;
• Analytical summaries and executive insights.

These outputs are generated by combining your uploaded data with statistical models and AI algorithms. They are designed to support your decision-making process — not replace it.

Limitations of AI

AI technology, while powerful, has well-documented limitations that are important to understand:

Historical Data Dependence

AI models learn from historical patterns. They cannot account for unprecedented events, sudden market shifts, regulatory changes, or other factors not represented in past data. For example, a demand forecast trained on pre-pandemic data would not have predicted COVID-era supply chain disruptions.

Data Quality Sensitivity

The accuracy of AI outputs is directly proportional to the quality of your input data. Incomplete records, data entry errors, inconsistent formatting, or unrepresentative time periods (such as holiday seasons or promotional periods used as "normal" baselines) will affect output reliability.

Uncertainty in Predictions

All forecasts involve uncertainty. Our models provide confidence intervals where possible, but actual outcomes may fall outside predicted ranges. Forecasts become less reliable as the prediction horizon extends further into the future.

Contextual Limitations

AI does not have access to information beyond what you provide and what is available through our data sources. It may not know about your specific business relationships, contractual obligations, regulatory requirements, seasonal patterns unique to your industry, or local market conditions that affect supply chain decisions.

Model Evolution

Our AI models are continuously improved and updated. This means that outputs generated at different times may vary, even with the same input data, as model refinements are applied. We aim to communicate significant methodology changes through product announcements.

Informational Purposes Only

All AI-generated content on SupplyChainStack is provided for informational purposes only. It is not intended to serve as:

• Professional supply chain consulting advice;
• A substitute for independent analysis by qualified professionals;
• A guarantee of future performance, market conditions, or supplier reliability;
• The sole basis for purchasing, stocking, sourcing, or logistics decisions;
• Legal, financial, or regulatory compliance guidance.

Supply Chain Decisions Involve Risk

Supply chain management inherently involves business risk. Decisions about inventory levels, supplier selection, freight routing, and cost management carry financial implications. AI-generated recommendations should be evaluated alongside your team's domain expertise, market knowledge, and business context before implementation.

Factors that AI may not fully account for include:

• Geopolitical events and trade policy changes;
• Natural disasters and weather disruptions;
• Supplier financial health changes;
• Currency fluctuations and commodity price volatility;
• Regulatory changes in import/export requirements;
• Competitive landscape shifts;
• Customer relationship dynamics and negotiation context.

Verify Before Acting

We recommend the following verification practices before acting on AI-generated content:

• Cross-reference AI recommendations with your own market intelligence and supplier relationships;
• Review forecast confidence intervals and understand the range of possible outcomes;
• Consult relevant team members who have on-the-ground knowledge of specific products, routes, or markets;
• Test incrementally — implement AI-driven changes gradually rather than all at once;
• Monitor outcomes — track actual results against predictions and adjust your reliance accordingly.

Professional Judgment

AI outputs are a tool to enhance — not replace — professional judgment. Experienced supply chain professionals should evaluate AI recommendations in the context of their industry expertise, business relationships, and strategic objectives.

If a recommendation from SupplyChainStack conflicts with your professional experience or seems inconsistent with market conditions you observe, trust your expertise. AI provides one data point among many that should inform your decisions.

Accuracy and Hallucinations

This phenomenon, sometimes called "hallucination" in AI research, means that AI models may:

• Present incorrect statistics or calculations as if they were accurate;
• Suggest suppliers, routes, or strategies that may not be optimal for your specific situation;
• Generate analyses that sound reasonable but are based on flawed pattern recognition;
• Produce inconsistent results across different runs with similar inputs.

We implement multiple layers of validation, including statistical checks and output filtering, to reduce the frequency of inaccurate outputs. However, no AI system is infallible, and critical business decisions should always involve human review.

Liability

Steeled Inc. does not warrant the accuracy, completeness, reliability, or timeliness of any AI-generated content. To the fullest extent permitted by applicable law, Steeled Inc. shall not be liable for any direct, indirect, incidental, consequential, or punitive damages arising from:

• Reliance on AI-generated forecasts, recommendations, or analyses;
• Business decisions made based on AI outputs;
• Inaccurate or incomplete AI-generated content;
• Differences between AI predictions and actual outcomes;
• Losses resulting from supply chain decisions informed by the Service.

Users assume all responsibility for evaluating the suitability and accuracy of AI-generated content for their specific use case. See our Terms of Service for complete liability terms.

Data Quality and Your Responsibility

The principle of "garbage in, garbage out" applies to AI systems. You are responsible for:

• Uploading accurate, complete, and representative data;
• Ensuring your data covers a sufficient time period to enable meaningful analysis;
• Flagging known anomalies in your data (e.g., one-time bulk orders, promotional periods, stockout periods);
• Regularly updating your data to reflect current business conditions;
• Understanding that outdated or incomplete data will produce less reliable outputs.

Per-Feature Notes

The following notes apply to specific features within SupplyChainStack:

Demand Forecasting

Forecasts are generated using statistical models (moving averages, trend analysis) combined with AI pattern recognition. Forecast accuracy depends on the volume and consistency of historical data. Products with limited sales history or highly variable demand patterns will produce less reliable forecasts. Confidence intervals are provided to indicate the range of uncertainty.

Inventory Optimization

Inventory health assessments classify products by risk level (healthy, warning, critical) based on sales velocity, trend analysis, and days since last sale. These classifications are algorithmic assessments, not absolute judgments. A product flagged as "critical" may still have business value depending on factors the algorithm cannot assess, such as contractual obligations or strategic customer relationships.

Supplier Evaluation

Supplier risk scores are based on available data including delivery history, pricing trends, and reported certifications. Scores do not account for private information about supplier financial health, management changes, or pending litigation. Always conduct independent due diligence before making supplier decisions.

Cost Analysis

Landed cost calculations and cost optimization suggestions are based on the data you provide and publicly available reference rates. Actual costs may vary due to negotiated rates, volume discounts, currency fluctuations, and regulatory changes. Cost estimates should be verified against actual quotes and invoices.

Risk Assessment

Risk alerts are generated based on data patterns and known risk factors. They are probabilistic assessments, not predictions of certain outcomes. The absence of a risk alert does not guarantee the absence of risk. Risk assessments should be supplemented with your own monitoring of geopolitical events, weather patterns, and industry-specific factors.

Marketplace Match Scores

Match scores between buyers and service providers are generated algorithmically based on stated requirements, geographic proximity, certifications, and service capabilities. Match scores reflect compatibility based on available data and do not constitute endorsements or guarantees of service quality. Provider credentials displayed on the Marketplace are self-reported and not independently verified by SupplyChainStack unless explicitly stated.

Freight Rate Estimates

Freight rate estimates are based on benchmark data and historical rate information. Actual freight rates vary significantly based on carrier negotiations, fuel surcharges, seasonal demand, lane-specific capacity, and current market conditions. Freight estimates are indicative only and should be confirmed with actual carrier quotes before making logistics decisions.

Best Practices for Using AI-Generated Content

1. Use AI as a starting point, not a final answer. Let forecasts and recommendations inform your research, then apply human judgment.
2. Maintain human oversight for all consequential decisions. AI should augment your team, not replace critical thinking.
3. Track prediction accuracy over time. Compare AI forecasts to actual outcomes to calibrate your level of reliance.
4. Keep data current. Regularly upload fresh data to improve the relevance and accuracy of AI outputs.
5. Diversify your information sources. Don't rely solely on SupplyChainStack for supply chain intelligence — use industry reports, supplier communications, and market research alongside our platform.
6. Escalate when uncertain. If an AI recommendation seems unusual or high-stakes, consult with experienced professionals before acting.
7. Document your decisions. When using AI-generated insights to justify business decisions, document the full context including what other factors you considered.

Contact Us

If you have questions about our AI technology, the methodology behind specific outputs, or this disclaimer, please contact us:

Email: supplychainstack@polsia.app

Entity: Steeled Inc., a Delaware Corporation

We welcome feedback about AI output quality and are committed to continuously improving the accuracy and usefulness of our platform.