Our AI principles
- Honest. AI drafts; humans approve.
- Respect platform rules. We never post where the directory doesn't permit third parties to.
- No fabrication. We never invent reviews, ratings, or customer feedback.
- Transparent. Any AI-generated content is marked as a draft in the app until a human approves it.
Where we use AI
- Reply drafting for reviews - you review and approve before anything goes out.
- Sentiment analysis and topic clustering across reviews.
- Listing extraction from public directory pages.
- Message personalization for review requests and follow-ups.
- Chatbot answers trained on your own site content.
Publishing rules
We publish AI-drafted replies directly only for Google Business Profile, and only when you have connected your Google account and clicked "Approve & post." For every other directory (Yelp, Facebook, Angi, HomeAdvisor, Houzz, BBB, YellowPages, BuildZoom, Zillow, Healthgrades, and others), RHAWK copies the reply to your clipboard and deep-links you to the directory so you paste and post it. This is not a technical limitation - it's the correct behavior under those platforms' terms.
What we will never do
- Create fake reviews or ratings.
- Post to a directory when the platform prohibits third-party posting.
- Impersonate a customer in messages.
- Send review requests to contacts you haven't uploaded consent for.
Models & providers
We use foundation models from Google (Gemini), OpenAI (GPT family), and, when appropriate, open-weight models. The specific model routed depends on the task (drafting vs. extraction vs. summarization). All providers are contractually barred from training on our customers' data.
Your data is not used to train foundation models
We do not permit any of our AI providers to train their foundation models on your account data, uploaded contacts, review content, or messages.
Human review
Every AI reply draft is marked as "Draft" until a human user clicks Approve. Automations (e.g. auto-approve high-confidence replies for 5-star reviews) are opt-in and can be turned off at any time from the settings page.
Report an AI concern
Spotted an AI-generated reply that looks wrong, unfair, or inappropriate? Email ai@ratehawk.app. We investigate every report.