If your tutor profiles list units or topics (equilibrium, electromagnetism, cellular respiration, organic mechanisms) the bot reads them and matches accurately. If profiles only say 'AP Chem specialist', unit-level matching is weaker but subject-level matching still works. Most services that care about unit precision add a topics list per tutor; the bot then routes the unit-specific question to the tutor who covers it best.

Yes, if your roster includes MCAT, GRE, DAT, or PCAT specialists. The bot reads the profiles and routes accordingly. Most services run a separate landing page for graduate exam prep because the price point and conversation pattern is different from undergrad coursework. The bot adapts when the visitor enters from that page, with longer-arc package recommendations rather than single sessions.

Yes. AP in May, SAT Subject Tests (where still relevant), university finals in December and May, MCAT testing windows, GRE rolling. The bot recognises these signals and recommends frequency accordingly. For a 2-week test prep window, two sessions. For a 5-month AP runway, weekly with a check-in cadence. Recommendations are grounded in tutor availability, not aspirational scheduling.

Yes, if those rates are on the pricing page. The bot quotes trial-session discounts, hourly rates, 4-packs, 10-packs, and exam-prep packages from real data. It will not invent discounts; if you want it to honour a referral code, add the rule to the prompt. Otherwise it quotes only what is published so refunds and accounting stay clean.

If lab content is documented (videos, demo writeups, lab report templates) the bot reads them and explains lab procedures, write-up format, and common lab questions. For services that bundle lab help with subject tutoring, this is one of the highest-value features because lab reports are often where students lose grade points despite understanding the material. The bot's job is to surface that help, not run virtual labs.

Yes. If your tutors list university courses they have TA'd or tutored (CHEM 112 at State U, PHY 121 at City Tech), the bot matches by course code when students mention it. This is especially valuable for university-adjacent services because course codes are how students think about their needs, not generic 'physics' framing. Most services tune the prompt to ask for course code if the student mentions university.

Logs stay in your WordPress database. Bring-your-own-key sends prompts directly to OpenAI, Anthropic, or Google with no third-party processor in between. That matters for students sharing grades, struggles, or coursework details: you control retention and redaction, and the plugin does not add a new vendor between you and the model. Apply your existing WordPress security to the log table.

Some homework help is appropriate; full homework completion is not. Most services tune the prompt so the bot can nudge a stuck student with a hint or a question, but routes longer help to a booked session. This protects the service's revenue model and the student's learning. The line is usually around 5 minutes: anything quick is a nudge, anything substantial is a session.