Skip to content
Stedle AI

Concept build

AI Front Desk for Dental Clinics

Paste a clinic’s website and, in about a minute, a team of agents turns it into a voice and chat front desk that answers pricing, services, hours and booking questions from the clinic’s own content.

AI Agents & AssistantsAI Workflow AutomationAI Integration & RAG EngineeringClinics & healthcare practices
Redrawn product screen. Names, businesses, people and prices are fictional sample data.

Verified in the code

evaluation questions passed
14/14evaluation questions passed
hallucination probes passed
5/5hallucination probes passed
average answer time
~4.4 saverage answer time
specialist agents
4specialist agents

The problem

Dental front desks spend the day answering the same questions about treatments, prices, hours and appointments, and patients calling after hours get nothing. An assistant here must never invent a price or a doctor.

What we built

One URL starts everything: the system crawls the clinic’s site, extracts structured facts with an LLM and builds a searchable knowledge base. An orchestrator classifies each question and calls specialist agents for pricing, services, general info and booking in parallel, then merges their answers into one reply, by voice or text.

Capabilities

What it does

  • One-URL setup: crawls up to 80 pages and extracts clinic details, services, doctors, pricing, hours, FAQs and policies
  • An orchestrator routes each question to specialist agents that run in parallel, then merges their answers
  • Grounding guard: when no relevant evidence is found, it tells the patient to contact the clinic instead of guessing
  • Booking agent gathers details over several messages, asking only for what is missing; dates like “next Monday” are resolved by code
  • Speech-to-text, text-to-speech and embeddings run locally; only reasoning steps call a cloud model
  • Every endpoint, including agent-to-agent calls, requires an API key
  • Tested on the public websites of two different practices without clinic-specific code

Under the hood

Orchestration

  1. 1Patient question
  2. 2Intent classifier
  3. 3Pricing · Services · General · Booking agents (parallel)
  4. 4Grounding check
  5. 5Answer merge
  6. 6Voice or chat reply

Stack

n8nOpenAI gpt-4o-miniQdrantPostgreSQLOllama embeddingsfaster-whisperPiper TTSFastAPIDocker
Next projectBilingual Pricing AssistantView case study

Want something like this for your business?

Book a free call. We’ll look at how your business runs and tell you honestly what’s worth automating, and what isn’t.

Book a callSend a message