Build Your AI Transparency Notice

The more detail you give us, the more useful your draft will be. Leave a question blank if it doesn't apply -- we'll simply omit that section rather than guess.

Your email address *

We'll email your draft AI Transparency Notice here as soon as it's ready. We never publish or share this address.

The Basics

Company name *

Product or feature name *

In a sentence or two, what does this product/feature do for the user or customer? *

Who is this notice for? (select all that apply)

End customers Customers' end users Regulators Other

Is this a first draft or an update to an existing notice?

First draft Updating an existing notice

Purpose and Features

What does the AI system do, and what outcome is it built to support? *

List the specific AI-powered features/capabilities in this product, if there are several

Data Sources and Storage

What categories of data does the AI touch? (select all that apply)

If "Other," please describe

Where does this data come from? (e.g., customer-provided data, user-submitted input, licensed/public data) *

Cloud provider(s) and physical storage region

Encryption approach (at rest / in transit)

How is customer data segregated from other customers' data, if at all?

Data retention and deletion practice

AI Model

Is the underlying model generative or closed/non-generative?

Generative (can produce novel text/images/etc.) Closed/non-generative (classifies, extracts, or scores against fixed categories) Not sure

Is the model built in-house, fine-tuned from a third-party base model, or used as-is via a third-party API?

Built in-house Fine-tuned from a third-party base model Used as-is via a third-party API Not sure

Model Training

Does your company train or fine-tune any AI models (in-house or on customer data)?

Yes No Not sure

If yes, what data is used for training/fine-tuning?

Is customer data ever used to train models?

Never Only with explicit customer authorization Yes, as a standard practice Not applicable

How are training environments secured and later decommissioned?

What change control exists before a newly trained model is released?

Principles, Bias, and Limitations

Any responsible data use principles worth stating (e.g., data minimization, no resale, no cross-customer leakage)?

Does this AI make or inform decisions about people (e.g., approval/denial, scoring, ranking, identity verification)?

Yes No

If yes, how is bias assessed or mitigated, if at all? (It is fine to say this work is planned or not yet formalized.)

What known limitations, failure modes, or edge cases should customers be aware of? *

Human Oversight and Risk

Is there a human review step anywhere in the process?

Yes No

If yes, where in the pipeline does review happen, and what triggers it?

What risks does your company actively manage/guarantee? (e.g., data storage integrity, segregation, uptime)

What is explicitly the customer/deployer's responsibility? (e.g., credential security, data accuracy, access management)

Customer Controls and Continuous Improvement

Can the customer/user opt out of or disable this AI feature?

Yes No Partially / in some cases

What other controls does the customer/user have? (e.g., correcting/flagging outputs, requesting human review, exercising data subject rights)

How do you keep the system and this notice up to date?

Contact and Links

What email address or team should customers contact to report errors or safety/security concerns with this AI system? *

Link to your Privacy Policy (optional)

Link to your SOC 2 / Trust Center page (optional)

Link to an AI Use / Vendor Management Policy summary (optional)

By submitting this questionnaire, you agree to our Privacy Policy.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

AI Transparency Notice Questionnaire