AI Transparency
AI at Elentra
Last updated November 9, 2026
Elentra’s AI features help faculty and staff work with content in Elentra, such as reviewing and improving narrative evaluations. They suggest; people decide. This page explains what our AI features do, what data they use, and how we protect that data. It is part of the AI Documentation referred to in our AI Feature Terms. If anything on this page conflicts with our AI Feature Terms or your agreement with Elentra, those terms control.
Our commitments
- Off until you turn it on. Every AI feature is disabled until your institution’s administrator enables it, and each feature has its own switch.
- AI is clearly labeled. We label AI features in Elentra so your users know when they’re using one.
- People stay in control. Features flag possible issues and propose changes. Nothing is changed until a user accepts or edits a suggestion.
- No training on your data. We do not use your data to train AI models, and our agreements with our AI providers prohibit them from doing so.
- Limited retention by AI providers. Our agreements with our AI providers allow them to keep your content only briefly, to process requests and make features faster, and never longer than 30 days, unless the law requires otherwise or they need to investigate possible misuse flagged by their safety systems. Those agreements prohibit them from using it for anything else or sharing it with their other customers.
- Not for high-stakes decisions on their own. AI features can help people evaluate performance, but they are not designed to grade, assess competence, or make admission, progression or similar decisions about anyone on their own. A qualified person at your institution is responsible for making the decision.
What our AI features do
Our AI features help people work with content in Elentra. For example, they can help reviewers check and improve narrative evaluations by flagging possible issues, such as spelling and grammar errors, unclear or ambiguous wording, inconsistent formatting or tone, details that don’t match the student a narrative is about, and language that may reflect bias, and by suggesting changes. Each suggestion shows what would change before it’s applied. Flags and suggestions are prompts for a person to review, not findings. We add and improve features over time. The commitments on this page and the data practices described in “How your data is handled” apply to all of them and are what you can rely on.
How your data is handled
- What is sent to the AI model: Only the content a user selects, plus the details from Elentra that a feature needs to do its job, such as the learner’s name.
- No training: Neither we nor our AI providers will use your data, including what you submit to AI features and what they generate, to train AI models.
- Once you accept or edit: Content you save after accepting or editing a suggestion becomes part of your records in Elentra, like anything else you enter, and is protected the same way.
- AI providers: Our AI providers are listed on our sub-processor list, which is available on request at info@elentra.com. They are bound to data protection terms consistent with your agreement with us.
- How we improve features: We measure how AI features are used, for example whether suggestions are accepted, skipped or edited and the reasons users give for skipping. We also review content to improve quality, but only after it has been de-identified so that it cannot reasonably be used to identify your institution or any individual. We never use any of this to train AI models.
Where content is processed
We use commercially reasonable efforts to process AI content in the same region where your Elentra production data is stored, where the AI model is available there. If it isn’t, we use another region with comparable legal protections, listed below. We never process AI content in countries identified as countries of concern under U.S. data security rules or subject to comprehensive U.S. sanctions.
| Where your production data is stored | Where AI features process content |
|---|---|
| United States | United States |
| Canada | Canada |
Controls and responsible use
- Administrators decide. Each AI feature is turned on or off in your Elentra product’s administrative settings. Before turning on the first feature, an administrator accepts our AI Feature Terms on behalf of your institution. New AI features also start turned off, including when they arrive in a product update.
- Keep patient information out. Don’t include protected health information, personal health information or anything that identifies a patient, or any payment card data or government ID numbers. Our AI features are not designed to detect or remove this information.
- Review every suggestion. AI can be wrong, incomplete or biased, even when it sounds confident.
- Keep a person in high-stakes decisions. Don’t rely on AI output alone for grading, assessing competence, progression, admission or similar decisions. A qualified person should review the information and make the decision.
Changes to this page
We update this page as our AI features change. If a change materially reduces your protections under our AI Feature Terms, such as a change to where content is processed, we’ll give at least 30 days’ notice by email to your designated administrative contact and in Elentra before it takes effect.