Legal

AI Disclosure

How AI is used in TheAutoReply.

AI Disclosure — TheAutoReply

Last updated: May 2026. To be reviewed by counsel before public deployment.

What this document is

TheAutoReply is an AI-powered service. We believe in being clear about that, both with our customers (the businesses who pay us) and with the public who reads the reviews and replies on Google.

This document explains: 1. How AI is used in our product 2. What our customers should know 3. What end-users (the people who write the reviews) should know 4. Your rights and our commitments

How AI is used

When you connect your Google Business Profile to TheAutoReply, our service:

  1. Reads new reviews posted to your profile, using Google's Business Profile API.
  2. Drafts a reply to each review using a large language model (Claude, by Anthropic), guided by your configured brand voice and our internal review-handling policies.
  3. Posts the drafted reply to Google as a reply from your business — either automatically (in auto-post mode) or after your approval (in approval mode).

The drafted reply is generated by AI. The reply is posted in your business's name, just as if you had typed it yourself. Google identifies the reply as coming from "the owner" — the same way it would label any reply you posted manually.

What our customers should know

When you sign up for TheAutoReply, you're authorizing us to:

  • Read reviews posted to your Google Business Profile
  • Draft replies on your behalf using AI
  • Post those replies to Google in your business's name (in auto-post mode)

You are the publisher. Replies posted through TheAutoReply are published under your business identity and are your responsibility, the same way replies you typed yourself would be. We are a tool. You are the speaker.

You can review every reply before it posts. Approve-by-email mode lets you see every drafted reply before anything is sent to Google. We recommend new customers start here for the first 1–2 weeks.

You can edit any reply. Even in auto-post mode, you can edit a reply for up to 30 days after it's posted, and the change syncs to Google.

You can pause or cancel anytime. One click in your dashboard.

You're responsible for the truth of any factual claims. If a reviewer says "I waited 50 minutes" and our AI replies "we'll fix our scheduling," you are agreeing on the truth of that claim. If you don't want to make that admission publicly, switch to approve-mode and edit the reply.

What end-users (reviewers and the public) should know

Replies posted through TheAutoReply appear on Google under your business's name. Google's interface labels them as coming from "the owner" — which is accurate, because you are the legal speaker even if our AI drafted the reply.

We do not impersonate individuals. Replies are signed in the business's name, not in the name of a fake first-name human. If a reviewer asks "is this an AI reply?", the business owner is responsible for answering honestly.

We don't process the reviewer's personal data. We read what the reviewer publicly posted on Google. We don't enrich it with other personal information. We don't track the reviewer across the web. We don't sell reviewer data — to anyone, ever.

What our AI cannot do, by design

  • It cannot dispute a fact stated in the review
  • It cannot accuse a reviewer of lying
  • It cannot threaten legal action
  • It cannot reveal private information (medical details, personal information, financial details)
  • It cannot post a reply containing customer-identifying information beyond what the reviewer themselves disclosed
  • It cannot use language flagged as toxic, harassing, or discriminatory

These guardrails are enforced at the prompt and post-processing layers. They are not optional features — they are baked into the product.

Reviews we will not auto-reply to

Some reviews require human judgment. When these patterns are detected, the AI flags the review for owner review and does not auto-post:

  • Allegations of physical injury, illness, or harm
  • Allegations of theft, fraud, or illegal activity
  • Mentions of pending legal action, lawyers, or attorneys
  • Reviews longer than 500 words
  • Reviews referencing specific staff members in a serious complaint

In these cases, the customer receives a same-day email asking them to review the situation themselves before any reply is posted.

Your rights

  • You can request a copy of every reply we have posted on your behalf — exportable as CSV.
  • You can request deletion of all your data — fully purged from our systems within 30 days.
  • You can transfer your account to another owner of your business at any time.
  • You can cancel at any time, with no penalty.

To exercise any of these rights, email us at [email protected].

How we handle the AI itself

  • We do not train our AI on your reviews or replies. Your data is yours.
  • We do not share customer data between accounts to "improve" individual replies. Each customer's brand voice is isolated.
  • We log AI generations for 90 days for audit and quality purposes, after which logs are deleted.
  • We use Anthropic's Claude as our model provider. Anthropic's data handling commitments are described at anthropic.com/privacy.

Updates to this disclosure

If we materially change how AI is used in our product, we will notify customers by email at least 30 days in advance. Material changes include:

  • Adding new ways AI generates content on your behalf
  • Changing what data is sent to the AI model
  • Switching AI model providers in a way that affects data handling

Non-material updates (typo fixes, new examples) are made silently with a "Last updated" date change.

Contact

For questions about this disclosure or how AI is used in your account:

Email: [email protected] Mail: FrontDesk Global, [LLC address to be added]

This is a v0 draft. To be finalized after counsel review prior to public deployment.