AI Policy

This AI Policy explains how National Foreclosure Prevention Group (“NFPGRP,” “we,” “us”) uses artificial intelligence in our services, what our AI tools can and cannot do, and how you should treat any output they produce.

1. Where AI shows up in our services

• Document scanner. When you upload a court filing or a servicer letter to your case, an AI model reads the document to pull structured fields (case number, index ID, court, hearing dates). We use Anthropic Claude Haiku for this because it is fast and reliable at structured extraction. The extracted fields are stored on your case record so our team can use them to look up your case directly with the county, or file a public-records (FOIL) request when a county does not have it online.
• AI Case Companion chat (Standard and Premium members). A conversational assistant that answers general educational questions about the foreclosure process, points you at relevant videos, forms, government resources, and attorney directories, and (in later releases) can reference the filings our team has already found for your case. Powered by Anthropic Claude Sonnet.
• Personal SMS AI companion (Premium members). Daily case update summaries and text conversations delivered by SMS, produced by the same underlying model as the in-app chat.
• Chat Access: AI Case Library (public, separate purchase). A conversational interface that helps you find public court forms and understand what each form is generally for. Powered by a large language model.

2. You are always talking to an AI when you use these tools

When you interact with the document scanner, the AI Case Companion chat, the SMS AI companion, or Chat Access, you are talking to a machine, not a person. Our team may review conversations for safety, quality, and support. We will tell you when a human is stepping in.

3. What our AI does not do

• Give legal advice. Our AI is not a lawyer. It cannot tell you whether a specific action is a good idea in your case, whether a specific defense will succeed, or whether a specific document will be accepted by your court.
• Represent you. Our AI does not appear in court, sign filings for you, or communicate with your lender on your behalf.
• Fill out forms for you. The AI Case Library supplies blank public forms and general context. You are responsible for filling them out, with an attorney if you can afford one.
• Predict specific outcomes. Timelines, credit-score projections, and case-status estimates are illustrative. They are not guarantees.

4. Accuracy and human review

Large language models can produce output that is confidently wrong (“hallucinations”), stale (not reflecting recent changes in law), or generic (not accounting for local rules). Before you act on anything our AI tells you, you should:

1. Verify it against the primary source (a statute, a form, a court's website)
2. Confirm it applies to your state, county, and situation
3. Consult a licensed attorney in your state if the stakes are significant

The document scanner is a convenience: check the extracted fields against the original document before you rely on them. NFPGRP is not responsible for decisions you make based on AI output without verifying it.

5. How your data is used with our AI

Conversations, documents, and prompts you provide are used to:

• Deliver the service you asked for in that session
• Improve our systems, in de-identified or aggregated form
• Respond to lawful requests or investigate abuse of our services

We do not sell your conversations. We do not use identifiable conversation content, case data, or uploaded documents to train third-party AI models. Our AI infrastructure provider (Anthropic) processes prompts under a zero-retention arrangement where available: they do not retain your content beyond what is necessary to return a response, and they do not use it to train their public models.

For a full picture of what we collect and how long we keep it, see our Privacy Policy and our Subprocessors list.

6. Safety and prohibited use

Do not use our AI tools to:

• Impersonate someone else, or submit information (including court documents) you do not have permission to share
• Attempt to extract training data, jailbreak safety filters, or bypass rate limits
• Give legal, financial, or medical advice to others while claiming it came from NFPGRP
• Automate abuse, harassment, or fraud

7. Age restriction

Our AI tools are for adults 18 and older. We do not knowingly provide AI services to minors.

8. State-specific disclosures

Some U.S. states require additional disclosures when a business uses AI in customer interactions (for example, California SB 942 watermarking rules and generative-AI training data disclosures, Utah's AI Policy Act, and others). Where such disclosures apply to a specific interaction, we will surface them at the point of use.

9. Changes to this policy

AI capabilities evolve fast. We update this policy as our tools change and as regulation catches up. Check the “Last updated” date at the top of this page.