No. SleekAI is configured with strict guardrails to decline any diagnostic, treatment, dosage, or symptom-interpretation question and refer the patient to a qualified clinician or emergency services. The system prompt explicitly enumerates declined topics and provides emergency-service language for urgency keywords. We recommend audit-testing the bot quarterly with edge-case prompts to confirm the guardrails hold.

HIPAA compliance depends on your overall stack and BAAs, not on a single plugin. SleekAI is a WordPress plugin that uses OpenAI's API; for US clinics handling PHI, you need a HIPAA-eligible OpenAI configuration, a BAA in place, encryption at rest on the WordPress database, and proper access controls. The plugin gives you the building blocks - data lives on your server under your control - but the compliance posture is a stack-wide responsibility.

Yes - it links to your existing scheduler (Calendly, Acuity, Doctolib, or a clinic-specific PMS) so the bot is the front door rather than a parallel booking engine. Patients see real availability from your calendar, the appointment lands in the system your staff already uses, and you avoid the double-booking risk that comes with shadow scheduling tools.

The bot is configured to recognise urgency keywords - chest pain, shortness of breath, severe bleeding, signs of stroke - and immediately direct the patient to call local emergency services rather than wait for chat triage. This is non-negotiable in the system prompt and tested as part of the standard configuration. We also recommend a visible disclaimer in the widget header so patients know the channel's scope upfront.

Yes - patients writing in Spanish, French, Arabic, Mandarin, or any common language get responses in their language, with operational details translated on the fly. For diverse practice areas this is genuinely useful: a non-English-speaking patient who wouldn't call the front desk will use a chat in their own language to book a routine appointment.

Yes, in your WordPress database under your control. Retention is configurable. For PHI we recommend short retention windows - days rather than years - encryption at rest on the database, role-based access on who can read the logs, and BAAs with every vendor in the stack including your hosting provider and OpenAI. Document the conversation log in your privacy notice and your records-retention policy.

Rescheduling and cancelling work through your existing scheduler's links - the bot doesn't directly modify calendar records. This keeps a single source of truth in your scheduling system and avoids race conditions between the bot's view and the actual calendar. Patients click through to the standard reschedule URL pre-filled with their booking ID.

Yes operationally - the bot accepts the booking like any other - but verification of the relationship and consent is a front-desk responsibility on arrival. We recommend the system prompt acknowledge third-party bookings politely and capture caller-and-patient names separately so staff have the right context. Sensitive consent and minor-patient rules stay with humans.