We built a free ROI calculator that runs the actual math on whether an AI receptionist beats your current setup. Here's a step-by-step walkthrough using a real 4-location dental group as the example.
The first question on every voice-agent discovery call is "what's this going to cost me?" and the honest second question is "is it actually going to pay back?"
To answer both at the same time, we built a free Cost Estimator on the site. No email gate to see the result, no marketing copy in the way. Plug in your numbers, get the 12-month TCO for in-house, outsourced, and AI-agent configurations side by side.
Below is a step-by-step walkthrough using a real customer scenario — a four-location dental group in the Mountain West with 1,800 inbound calls/month — so you can see exactly how the calculator works and what assumptions it's making.
I'm using a real composite of three dental customers we've worked with. The numbers are slightly rounded but the proportions are accurate.
Open the Cost Estimator and fill in:
The calculator immediately produces three scenarios.
Cost line items:
Revenue line items:
The calculator's verdict on in-house: a positive but leaky operation. Real revenue happens, but 225 calls/month vanish into voicemail and only 18% come back.
The calculator assumes a competitive blended answering service like Ruby or AnswerConnect at scale:
Cost line items:
The calculator surfaces a yellow warning here: at 1,800 calls × 3.5 min, the outsourced model is dramatically more expensive than in-house because the overage rate destroys the unit economics. For high-volume dental, the answering service is the worst of all three options.
Revenue line items:
Cost line items:
Revenue line items:
The calculator's headline output:
Scenario Annual cost Annual revenue Net contribution
───────────────────── ──────────── ────────────── ────────────────
In-house receptionists $76,200 $238,680 $162,480
Answering service $100,704 $188,496 $87,792
AI voice agent $32,052 $359,856 $327,804
The agent doesn't just save on labor cost. It adds $121,000 in revenue by capturing every call and converting at a higher rate. Net contribution improvement vs in-house: $165,324/yr.
Build cost: $9,950. Monthly net improvement vs in-house: $13,777. Breakeven: day 22.
Most customers don't believe that number until they see it in their own EHR after the first month. We had one dental group call us to ask if their PMS was double-counting — it wasn't, they'd genuinely added 180 booked appointments month one because the agent caught every single after-hours call from people in dental pain Googling "emergency dentist near me."
The calculator has a sensitivity slider. The two assumptions worth pressure-testing:
a) Booking rate uplift. If you don't believe the 7-point booking rate lift (42% → 49%), drag it down to a conservative 3 points. The agent scenario still beats in-house by $58,000/yr in net contribution.
b) Call value. If your appointments are worth less — say $180 instead of $340 — the agent still wins, but by a smaller $43,000/yr margin.
The only configuration where the agent loses to in-house: call volume below ~250/month AND average call value under $100. That's not a dental group. That's a single-operator coach business. For that customer, we recommend a chatbot, not a voice agent.
To be honest about what we're modeling and what we're not:
The calculator's output should be read as a floor, not a ceiling.
"Won't patients hate talking to a robot?"
This is the most common objection. The honest answer: a poorly-built agent — wrong voice, awkward prompts, no escalation path — gets hated. A well-built one gets a 4.6 / 5 rating from real customers in our post-call survey. Patients care that someone answers the phone and books their appointment correctly. The "human or AI" distinction matters less than founders think.
The other honest part: we always include a human escalation. Any caller who says "I want to talk to a person" gets routed to a real number. About 6% of calls trigger this. The rest are completely handled by the agent.
Run your own numbers — takes 2 minutes, no email required to see the result. If the math looks good, book a call and we'll quote a specific scope for your business inside 24 hours.
If the math doesn't look good — your volume is too low, your call value is too small, your current operation is actually well-tuned — I'll tell you that on the call. We turn down about 15% of inbound discovery calls because the math doesn't support an install. We'd rather tell you no than ship a pilot that won't pay back.
For a different shape of business, let me run the same calculator against an HVAC scenario so you see how the numbers shift.
Inputs:
In-house scenario:
AI agent scenario:
Net contribution delta: agent adds $90,584/yr in net contribution. Build cost: $9,950. Breakeven: day 40.
The pattern across verticals: the higher the per-booking value, the faster the breakeven. HVAC at $890/booking pays back faster than dental at $340/booking, which pays back faster than coffee-shop catering at $80/booking (which actually doesn't pay back — we'd recommend a chatbot, not a voice agent, for that one).
A common situation: the SMB owner doesn't actually know their current booking rate, or their missed-call percentage, or their average appointment value. That's a problem worth fixing before you buy an agent — but here are reasonable defaults if you need to estimate:
If you don't know any of these, the right first step isn't a voice agent — it's a 30-minute audit of your current call data. We'll do that for free on a discovery call.
A lot of "AI ROI calculators" are toys that compare a SaaS sticker price to a labor cost and call it done. Our calculator models four cost lines for the AI scenario that most don't:
Modeling all four gives you a real number to compare against, not a sales-page-fantasy number.