Yes. If your math sequence is published on the academics page, the bot reads each section and answers placement questions in context. Many private middle schools run three math sections in 6th (standard, honors, advanced) with measurable percentage splits, and the bot quotes those splits from the page rather than inventing them. For families considering accelerated placement, the bot describes the on-campus placement test and routes appeals to the math department, not to general admissions.

Yes, at the policy level. The bot quotes your published test-acceptance policy (ISEE Lower Level, SSAT Middle Level, or both), the school code, and the deadline for sending scores. It points families to ERB or SSAT for actual registration rather than collecting test fees in chat, which keeps the testing relationship between the family and the provider where it belongs.

Yes. The system prompt routes specific child questions (transcripts, accelerated placement appeals, recommendation letters) to the admissions director by email rather than answering them. The bot collects the parent's contact and the topic only, with no transcript or letter content stored in the chat log. Most schools consider this the right boundary because placement decisions involve teacher judgement the bot can't replicate.

By quoting your published aid policy verbatim and routing award-amount questions to the admissions or aid office. The bot can explain the SSS or PFS application process, the typical deadline (February 1 at most schools), and historical averages if you've published them ('roughly 28% of families receive aid; average award is 40% of tuition'). It never quotes a specific award for a specific family, which is the right line.

Yes. The bot reads your live published page each conversation. If you reschedule the November placement test to early December because of a weather closure, the bot follows the moment the page updates. No retraining, no model fine-tune. For long-form curriculum guides or course catalogs, OpenAI Files as a vector store handles up to 1GB of content with accurate retrieval per chunk.

Yes. Multibot lets you run one bot for grades 6-8 and a separate bot for grades 9-12, each with its own system prompt, presets, and display conditions. Display conditions can match URL pattern, post type, page parent, or user role, so the bot scoped to the middle-school section answers from middle-school content only. Schools that run a K-12 site often use this to keep the bot focused on the correct division.

FERPA applies to education records the school maintains. The bot answers from your published WordPress pages and does not create records. Conversation logs live in your WordPress database with the security you've configured, and many schools set short retention windows (30 to 90 days) on chatbot logs. Bring-your-own-key sends prompts directly to OpenAI, Anthropic, Google, or OpenRouter, with no third-party processor between you and the model.

Yes. Any language the underlying model supports. Spanish, Mandarin, Portuguese, French, Korean, Arabic, and many others all work natively. The bot reads your English-source pages and answers in the parent's language, which matters for international families considering relocation. WPML and Polylang sites can also map per-locale content so the bot pulls from the appropriately translated page when a parent is on a localised URL.