No. The system prompt explicitly forbids collecting child names, specific ages beyond a general band ('3-year-old', '4-year-old'), medical info, allergy details, or any other sensitive data. Parent contact (name and email) is only collected for tour bookings, and only after the parent volunteers it. Anything beyond that gets routed to the director instead of stored in the conversation log.

If your timeline, fees, and decision dates are on a WordPress page, the bot reads and quotes them. For nuanced admissions specifics, the bot routes to the director rather than guessing. This includes sibling-priority decisions, waitlist movements, and any policy that depends on the year's specific cohort, which the bot has no way to know.

Yes. The bot reads your About / Philosophy page and answers in your director's published voice. Reggio Emilia, Montessori, Waldorf, play-based, or hybrid, the framing matches what you publish. If your approach evolves or you publish a new philosophy statement, the bot reflects the change immediately, with no retraining required.

Yes. Hand off via webhook to Calendly, Acuity, or your custom tour form. The bot collects the parent's preferred slot in conversation, then either passes them to your booking link with the slot pre-selected or fires a webhook that creates the tour appointment server-side. Most preschools use deep links because the booking confirmation is cleaner for the parent.

Conversations are between the bot and adult parents on your school's WordPress site. Because the bot doesn't collect child data, COPPA isn't triggered. FERPA applies to education records, which the bot doesn't generate; it answers from your published pages only. Logs stay in your WordPress database with whatever security you've configured, and many schools set short retention windows on the conversation history.

Yes. Any language the underlying LLM supports. Spanish, Mandarin, Portuguese, Arabic, French, and 50+ others all work natively. The bot reads your English source pages and answers in the parent's language, which matters for schools serving immigrant or international communities. Tone and warmth carry across languages because the underlying models handle that competently.

If your ratios and staff bios are published, the bot quotes them. Most preschools publish the legal ratio plus a 'typical' lower ratio, and the bot will explain both honestly. For specific teacher assignments per child, that's a director conversation: too much depends on the cohort, the year, and operational realities the bot has no way to know without making something up.

Yes, at the policy level. If your food programme, nap schedule, and potty-training policy are on a WordPress page, the bot reads and explains them. For child-specific concerns ('my daughter is still in pull-ups, will that be a problem?'), the bot routes to the director rather than guessing. That's the right boundary: policy is published, individual circumstances are a tour conversation.