Bilingual AI Patient Agent

Patients text the practice at all hours asking about prices, hours, and appointments — and a busy front desk can't always answer fast. Slow replies lose bookings, and Spanish-speaking patients often wait even longer.

I built an AI assistant that answers patient texts automatically, around the clock, in both English and Spanish. It reads each message, looks up that patient's real history and the real open appointment times, books or cancels the appointment, and texts back — usually in under a minute. Patients feel like they're texting a very fast, very informed front desk.

1. A patient text comes in. Junk, reactions, and "stop" requests are filtered out first, so the AI only spends effort on real questions.
2. The AI pulls up the patient's file, checks their existing appointments, figures out their language, and notices if they're asking about price — all in a few seconds.
3. It uses a set of 12 connected tools to look things up for real instead of guessing: open appointment times, pricing, locations, and answers to common questions.
4. When a patient says yes to a time, it books the appointment straight into the practice's medical-records system and texts a full confirmation. When they ask to cancel, it cancels — and double-checks it really went through.
5. Every outgoing text passes one shared safety check first: has the patient opted out, is a staff member already handling this, are the dates real, and are we about to repeat ourselves.

Patients get accurate, personal answers in 10 to 25 seconds at any hour, the practice books appointments without staff lifting a finger, and the whole thing runs for roughly $60 to $120 a month in AI costs.