AI Transparency Policy

What This Policy Covers

This policy describes how Essent uses AI in its products, services, operations, and published content and how Essent communicates the role of AI to users.

This policy does not govern the independent practices of companies Essent does not own or control, including third-party AI providers and processors. Privacy and data-processing practices are addressed in the Essent Privacy Policy, applicable data-processing documentation, and provider terms.

Responsible AI Principles

Essent’s approach to AI is guided by transparency, reliability, fairness, security, customer control, appropriate human oversight, and accountability. These principles inform how Essent designs, evaluates, documents, and uses AI Features.

Use of AI

Essent uses AI systems, including large language models, reasoning models, machine-learning tools, retrieval systems, agent frameworks, workflow automation, and MCP Services, to support and enhance products, services, operations, and content.

AI may be used to:

• generate or assist in drafting content;
• summarize, transform, classify, or translate information;
• provide recommendations, explanations, or insights;
• retrieve information from authorized sources;
• automate repetitive or data-intensive tasks; and
• recommend, initiate, or execute actions when authorized and appropriately controlled.

AI is intended to augment human capabilities. It is not a substitute for appropriate human judgment, review, approval, or professional expertise.

AI-Assisted Public Content

Essent may use AI to help outline, draft, transform, translate, or improve publicly available content. Published work may combine human-authored and AI-assisted material.

Essent identifies public content that materially includes AI assistance through a notice, label, indicator, or similar disclosure where appropriate. A disclosure may appear in the article, page, document, application, or accompanying metadata.

Validation and Human Review

Public AI-assisted content is reviewed and edited before publication. Review may include checks for clarity, formatting, technical accuracy, relevance, inclusive language, and alignment with the intended audience. AI-generated code used in published examples or products is subject to testing appropriate to its purpose.

AI-generated information may be inaccurate, incomplete, outdated, misleading, biased, or fabricated. Users should evaluate important AI-generated information before relying on it. Contractual human-review obligations for customers are stated in the Essent AI Services Addendum.

AI Indicators and Disclosures

Essent may use indicators such as "AI Generated,” "AI Assisted,” or comparable notices to help users understand when AI materially contributed to content or functionality. The form and placement of an indicator may vary by product, medium, use case, and applicable legal requirement.

An AI indicator describes the role of AI but does not itself guarantee the accuracy, originality, completeness, or suitability of the associated content.

AI Recommendations and Actions

Some AI Features may retrieve information, provide recommendations, draft communications, create workflows, or interact with connected systems. Essent designs these capabilities around authorization, access controls, customer configuration, and differentiated levels of action.

Customers remain responsible for configuring user access, connected data sources, exposed tools, permissions, approval requirements, and oversight appropriate to their deployment. Detailed contractual requirements for agentic actions and human review are contained in the AI Services Addendum.

AI Models and Providers

Essent may use multiple AI model types and providers. Models, versions, hosting arrangements, routing, and default selections may change as services evolve.

This policy does not list every provider or model. Information concerning personal-information processing appears in the Privacy Policy, and provider-related contractual obligations appear in the AI Services Addendum.

Customer Data and Model Training

Essent’s contractual position is that customer Inputs, Outputs, and Customer Data are not used by Essent to train or improve shared or generally available AI models unless the customer affirmatively opts in. Additional details are provided in the AI Services Addendum and Privacy Policy.

Feedback and Issue Reporting

Essent may provide channels through which users can report inaccurate, harmful, misleading, abusive, or otherwise problematic AI behavior or content. Reports may be reviewed to investigate issues, improve safeguards, support users, and enforce applicable terms.

Changes to this Policy

Essent’s use of AI may evolve. Essent may update this policy to reflect changes in products, services, practices, technology, or law. Changes will be posted on this page or otherwise communicated as required.

This draft page was created with the assistance of AI and reviewed as part of the Essent policy development process.