Automate Patient Scheduling and Follow-Ups with Custom AI

Most dental practices use tools like Weave or Lighthouse 360 for reminders. These services send templated messages but cannot handle conversational replies. If a patient responds, “I need to move to next week,” or “Is my appointment with Dr. Evans or Dr. Smith?,” the office staff must manually intervene. The automation is a dead end, creating a hidden workload of constant monitoring and manual responses.

A busy 4-person front office team sends 40 reminders a day. About 15% of patients reply with a question or reschedule request. This means 6 conversations per day that pull staff away from in-person patients to log into the PMS, find an open slot, and engage in a slow SMS back-and-forth. Each reschedule consumes 5-10 minutes of focused time.

Furthermore, the recall systems built into a PMS like Eaglesoft are too rigid. They send reminders based on a fixed 6-month schedule but cannot manage custom follow-up timelines for crowns or implants. Staff resort to spreadsheets and manual calendar entries, which inevitably leads to patients falling through the cracks and lost revenue.

Syntora would start by auditing your current patient communication workflows and connecting to your Practice Management System (PMS). For modern cloud platforms like Open Dental or Dentrix Ascend, we would use their official API. For older on-premise systems, we would design and implement a secure data connector that monitors the database for relevant changes or parses specific email notifications. The system would pull the next 14 days of scheduled appointments and your current patient waitlist into a Supabase database for real-time access.

We would then design a conversational AI agent using the Claude API, which would be trained on your specific office policies: cancellation rules, new patient protocols, and provider-specific procedures. When a patient texts to reschedule, the agent would check the PMS for open slots that match the original appointment's duration and provider, then offer valid options for the patient to choose from. We have built document processing pipelines using Claude API for financial documents, and the same pattern applies to structuring and responding to patient communications here.

The core system architecture would be a set of Python functions deployed on AWS Lambda. A scheduled event would trigger confirmation texts 48 hours before appointments via Twilio. Inbound SMS from a patient’s reply would trigger the AI agent for processing. A cancellation event detected in your PMS would immediately trigger another function that texts patients on your waitlist. The first patient to confirm would claim the slot, and the system would write the change back to your PMS.

For intelligent follow-ups, the system could scan completed appointments for specific procedure codes. For example, a root canal (CDT code D3310) could automatically schedule a 2-week check-in text. All system actions and conversations would be logged with structlog for auditing. Your office manager would receive a daily summary via email showing appointments confirmed, rescheduled, and waitlist-filled, along with any system exceptions.

Typical build timelines for this complexity range from 6 to 10 weeks, depending on PMS integration complexity and desired feature set. The client would need to provide API access credentials or technical documentation for their PMS, clear documentation of office policies, and access to a Twilio account. Deliverables would include the deployed cloud infrastructure, source code, a Supabase instance populated with relevant data, and detailed documentation for ongoing maintenance.