Background

I am Mario Cuevas, founder of AutoMeit. I built this product after watching aesthetic practices and home-services businesses burn through six-figure marketing budgets only to lose the resulting calls to voicemail. The pattern was the same everywhere: ten thousand a month on Google Ads and Instagram, a $600 consult ringing through to a busy front desk at 2 PM, voicemail, no callback, and a competitor down the street booked the patient by 3.

My background is in AI and voice systems. I spent years building speech and language pipelines, evaluating model behavior, and shipping production systems where latency and accuracy mattered. AutoMeit is the application of that work to a problem I watched cost real owners real money. The choice of medical spas as the initial customer profile was deliberate. Aesthetic practices have measurable ticket values, deep treatment menus, predictable call patterns, and a documented gap between marketing spend and answered phones. Every input to the math is observable, which means the ROI of an AI receptionist is provable rather than vibes-based.

Why Med Spas as the ICP

Medical spas are an unusual market. The average new-patient ticket is $200 to $500, the average lifetime value is $2,000 to $5,000, and the typical practice misses 22 to 28 percent of inbound calls. Roughly a third of inbound volume arrives outside business hours when no human is at the desk at all. That combination of high ticket value, high call volume, and high miss rate produces a documented annual loss of $50,000 to $200,000 per location. Fixing the answer rate is not a productivity exercise. It is the largest single revenue lever a single-location practice has access to without spending a dollar more on marketing.

The current generation of answering services and IVR systems do not solve this problem. Live virtual receptionists charge per minute, do not know the practice's treatment menu, and cannot book directly into Boulevard or Mangomint. Auto-attendants route callers through phone-tree menus that 60 percent of patients abandon. The opportunity sits in the middle: a voice agent that handles a Botox consult the way a trained med spa coordinator would, books straight into the calendar while the caller is on the line, and costs less than two recovered consultations a month to run.

How AutoMeit Was Built

AutoMeit's voice stack is built on three production-grade primitives. Deepgram handles speech-to-text. We selected Deepgram for its sub-300ms streaming transcription, its accuracy on aesthetic vocabulary (Botox, Dysport, hydrafacial, microneedling, GLP-1, semaglutide), and its WebSocket-based architecture which lets us pipeline transcription into language model inference without waiting for the caller to finish speaking. Gemini 2.5 Pro handles intent classification, treatment qualification, and dialogue policy. We tested Gemini against GPT-4o, Claude 3.5 Sonnet, and Llama 3.1 70B on a 200-call benchmark of real med spa intake recordings and Gemini won on a combination of cost, latency, and adherence to the qualification flow. ElevenLabs handles voice synthesis. ElevenLabs Turbo v2.5 was selected for its sub-400ms time-to-first-audio and its ability to clone a single voice persona across thousands of calls without drift.

Before any client onboards, the production pipeline is tested against 13 canonical med spa scenarios: first-time injectables consult, returning client rescheduling, Spanish-only caller asking for laser hair removal pricing, after-hours caller in distress about a botched Botox injection, package-bundle pricing question, GLP-1 weight-loss intake screening, no-show recovery callback, payment-plan financing question, treatment-room emergency interruption mid-call, mid-conversation language switch, intentionally hostile pricing-shopper, transfer-to-human request, and a "wrong number" hang-up that still needs to log a clean disposition. The current pipeline passes all 13. The outbound battle test (Aria placing reactivation calls to dormant patients) scored 4.0 out of 5 on a blind operator panel against three competing AI voice products. The bar is moving every release.

Connect

If you run a medical spa, an aesthetic practice, or a home-services trades business and you want to compare what your phone system is doing today against what an operator-built AI voice agent does, the easiest path is to book a 20-minute revenue audit. We pull your monthly call volume, miss rate, and average ticket, and run the math on the recoverable revenue your current setup is leaving on the floor.

You can also reach me directly at [email protected] or on LinkedIn. AutoMeit is run out of Atlanta and serves clients in all 50 states.