How we use AI

It always identifies itself as AI

Every conversation opens with Paige introducing itself as an AI assistant, in its very first message. Nobody talks to Paige believing they are talking to a person — consistent with California's bot disclosure law (SB 1001) and FTC guidance on AI transparency.

What Paige does

Paige answers general questions about the office, its team, and how the mortgage process works, using the content the office has published. When a visitor wants to go further, Paige collects their contact details — with their consent — and hands the conversation to a licensed loan officer at that office.

What Paige will never do

• Never quotes rates, APRs, or payments. Pricing a loan requires a licensed professional looking at a real file.
• Never approves, pre-approves, or evaluates eligibility. Paige is not an underwriter and does not pretend to be one.
• Never gives financial, legal, or tax advice.
• Never asks for — and will not keep — sensitive numbers. If a visitor types a Social Security number or a card number, it is automatically removed before the AI model ever sees it and before anything is stored.
• Never replaces a person. The goal of every conversation is a handoff to a licensed loan officer.

Guardrails behind the scenes

Conversations are limited in length and frequency to prevent abuse, and each office's AI usage is capped. The assistant runs on a commercial large-language-model provider under an agreement that does not permit training on this data. Transcripts attached to a lead are stored with the same protections as any other lead data — see our Privacy Policy.

Fair lending

Paige is instructed to treat every visitor equally and consistently with fair-lending law (ECOA, Fair Housing Act). It does not make — and is built not to make — lending decisions of any kind.

Questions

Write to contact@branchpage.com and a person will answer.