The hidden tax on every dental front desk
Dental front desks spend more time on insurance Q&A than any other single task. "Do you take Aetna PPO?" "What's my copay for a cleaning?" "Does my plan cover orthodontia?"
Multiply 40 of these per week by roughly 3 minutes per interaction across 50 weeks and you get 100 hours per year of front-desk time on a question that has a structured answer. That's 2.5 weeks of full-time work spent reciting your in-network list.
Why the chatbot version doesn't work
The instinct is to build a chatbot. "Just put a chatbot on the website that answers FAQs." Most practices have tried this. Most practices have given up because:
- Patients don't ask the question the chatbot expects. The chatbot has buttons for "insurance" but the patient types "do u guys take MetLife?" and the chatbot fails.
- The answer is more nuanced than a simple yes/no. "We take Delta Dental PPO but not Delta Dental DHMO" is the actual answer. Chatbots flatten that.
- The next question is more useful than the first. The patient who asks "do you take MetLife?" actually wants to know "what will my cleaning cost?" and you want them to ask that follow-up because you can answer it and book the appointment.
What an actual fix looks like
The right architecture isn't a chatbot. It's a structured in-network list paired with an agent that handles natural language:
- In-network plan list: maintained per practice. Specific carriers plus specific plan tiers (PPO vs HMO).
- Typical cost-share table: for common procedures (cleaning, exam, X-rays), what the patient typically pays out of pocket.
- Escalation criteria: when the agent doesn't know (complex pre-treatment estimate, plan-specific exclusions), it schedules a callback with the billing coordinator. It doesn't make up answers.
The agent takes the patient's natural-language question, maps it to the structured data, and gives a real answer:
"Yes, we're in-network with MetLife PPO. For a routine cleaning, most MetLife PPO patients pay around $30-45 out of pocket. Your exact amount depends on your remaining annual benefit. Want me to book you in for a cleaning so you can use this year's benefit before year-end?"
That's the response that converts. Not a yes/no. A real answer that segues into a booking off the live calendar.
What about complex cases?
No AI can answer every insurance question. Pre-treatment estimates for complex work (crowns, implants, orthodontia) require actually contacting the carrier. The right design is clear escalation: the agent knows what it knows, knows what it doesn't, and routes the doesn't-know cases to your billing coordinator with full context already captured.
The billing coordinator's day shifts from "answer 40 routine insurance questions" to "handle the 8 complex pre-treatment estimates that actually need expertise."
The HIPAA piece
Insurance Q&A involves PHI, specifically the connection between a patient's identity and their insurance plan. A dental AI receptionist has to handle that carefully: a BAA signed during onboarding, an outbound PHI guard that blocks before send and logs the block, and voice locked to the OpenAI engine because xAI offers no BAA covering call audio.
SOC 2 work is an audit in progress. We'll say "certified" when a report exists and not before.
What it does not do
It does not write into Dentrix, Eaglesoft, or Open Dental. That integration doesn't exist in this product. The agent works alongside your practice management system and books through RexRuby's own calendar, Acuity, Jane, Google Calendar, or Cal.com.
What we built
Rex handles insurance Q&A for dental practices using a structured in-network plan list plus a typical cost-share table, in 11 languages, across voice, SMS, and webchat. Ruby runs recall at the 45-day lapse window, no-show follow-up, first-visit check-ins, and NPS.
Dental plans are $999/mo flat, including the HIPAA module, document intake, and a signed BAA at onboarding. See the dental page