Whether an AI receptionist makes economic sense for your business comes down to four numbers: average ticket size, missed-call rate, close rate on recovered calls, and the AI's monthly cost. Here's the calculator math, with worked examples for HVAC, dental, and law firms — plus a printable break-even table you can use to size your own decision in 60 seconds. Most service businesses break even on call #4 of the month. The rest of the math is upside.
The formula
Annual recovered revenue =
monthly missed calls × close rate × average ticket × 12 months
Annual ROI =
(annual recovered revenue - annual AI cost) / annual AI cost
That's it. Four inputs, one calculation. The trap is using vanity numbers — "we miss like 5 calls a month" when your call logs say 35. Pull the actual data first. (If you want to see what the AI receptionist solutions cost side looks like, that page has the short version.)
The four inputs explained
1. Average ticket size
The average revenue from a closed job, case, or new patient. Use trailing 90-day data. Don't average lifetime value with first-visit revenue — they're different decisions.
- HVAC: average repair ticket, average install ticket, weight by mix
- Dental: average new-patient first-year value
- Law firm: average case fee or retainer
- Plumbing: average service-call ticket
- Salon: average new-client first-year value
2. Missed-call rate
Pull from your phone provider's call log. Count calls that went to voicemail, hit busy, or rang and weren't answered. Most service businesses underestimate this number by 2–4x. <!-- "Missed calls" includes no-answer, busy, and abandoned voicemail -->
Common ranges (calls per month missed):
- Solo operator: 30–80
- 5–10 person shop: 80–250
- Multi-location (50+ employees): 300–1,000+
3. Close rate on recovered calls
Of the missed calls the AI recovers (i.e., books or qualifies), what % become paying customers? Use the same close rate you'd apply to a normal inbound call. Be conservative — recovered calls aren't quite as warm as live-answered ones.
- HVAC service requests: 60–80%
- Dental new patients: 40–60%
- Law firm intakes: 25–45%
- Plumbing service requests: 65–85%
- Roofing leads: 15–30%
4. AI receptionist monthly cost
For most service businesses, this is $200–$400/month all-in (vendor + integrations + minutes). Use the high end for safety.
Worked example: HVAC
Inputs (typical solo HVAC operator, suburb market):
| Input | Value |
|---|---|
| Average ticket (repair) | $480 |
| Average ticket (install) | $7,200 |
| Mix-weighted ticket | $800 |
| Missed calls / month | 30 |
| Close rate on recovered | 70% |
| AI cost / month | $300 |
Math:
Monthly recovered revenue:
30 × 0.70 × $800 = $16,800
Annual recovered revenue:
$16,800 × 12 = $201,600
Annual AI cost:
$300 × 12 = $3,600
Annual ROI:
($201,600 - $3,600) / $3,600 = 5,500%
Break-even at: 0.5 calls per month recovered. The first call of the month pays for the year.
Worked example: Dental practice
Inputs (single-doctor practice, mixed insurance):
| Input | Value |
|---|---|
| Avg new-patient first-year value | $1,200 |
| Missed calls / month | 25 |
| Close rate on recovered (new patient) | 50% |
| AI cost / month | $350 |
Math:
Monthly recovered revenue:
25 × 0.50 × $1,200 = $15,000
Annual recovered revenue:
$15,000 × 12 = $180,000
Annual AI cost:
$350 × 12 = $4,200
Annual ROI:
($180,000 - $4,200) / $4,200 = 4,186%
Worth noting: dental new-patient lifetime value is often 5–10× the first-year value [VERIFY], so this is the conservative number.
Worked example: Law firm
Inputs (small PI / family law practice):
| Input | Value |
|---|---|
| Average case fee | $4,500 |
| Missed calls / month | 15 |
| Close rate on recovered | 30% |
| AI cost / month | $400 |
Math:
Monthly recovered revenue:
15 × 0.30 × $4,500 = $20,250
Annual recovered revenue:
$20,250 × 12 = $243,000
Annual AI cost:
$400 × 12 = $4,800
Annual ROI:
($243,000 - $4,800) / $4,800 = 4,963%
Even at a conservative 30% close rate (meaning 70% of recovered calls go nowhere), a single retained PI case typically pays for 5+ years of AI receptionist service.
Break-even thresholds
The break-even calc answers: how many calls do I need to recover per month for the AI to pay for itself?
Break-even monthly calls =
AI monthly cost / (close rate × average ticket)
At a $300/month AI cost and a 50% close rate, here's the break-even monthly call volume by ticket size:
| Average ticket | Break-even calls/month | Break-even calls/week |
|---|---|---|
| $100 | 6.0 | 1.5 |
| $200 | 3.0 | 0.7 |
| $400 | 1.5 | 0.4 |
| $600 | 1.0 | 0.3 |
| $800 | 0.75 | 0.2 |
| $1,000 | 0.60 | 0.15 |
| $1,500 | 0.40 | 0.1 |
| $2,500 | 0.24 | 0.06 |
| $5,000 | 0.12 | 0.03 |
At a $400/month AI cost and a 30% close rate (more conservative):
| Average ticket | Break-even calls/month |
|---|---|
| $100 | 13.3 |
| $250 | 5.3 |
| $500 | 2.7 |
| $1,000 | 1.3 |
| $2,500 | 0.5 |
| $5,000 | 0.3 |
| $10,000 | 0.13 |
If your average ticket is over $400 and you miss more than a few calls a week, the math is strongly positive — assuming the AI can actually recover those calls (it can, if configured well).
What the calculator misses
These factors aren't in the headline math but matter:
- Customer LTV. A new dental patient is worth $1,200 in year one and often $8,000+ over a decade [VERIFY]. The first-year number understates the impact.
- Reputation effects. Each missed call has a small chance of becoming a negative review ("they never answer"). AI removes this risk.
- Owner time saved. If the owner answers the phone today, the AI buys back hours that translate into more billable work or sleep.
- Outbound capacity. Some AI receptionists also handle outbound — appointment reminders, follow-up calls, win-back campaigns — that produce additional revenue not in the formula.
- Spam filtering. AI silently absorbs robocalls and spam, which is invisible time savings for whoever was handling those before.
When the math doesn't work
The AI receptionist isn't ROI-positive in every case. Three patterns where it can struggle:
- Average ticket under $100 with low close rates. A $50 retail call with a 20% close rate needs 30+ recovered calls/month to break even at $300/month AI cost.
- Already 95%+ pickup rate with a great human team. If your existing receptionist almost never misses a call, the upside is mostly after-hours coverage and overflow.
- Calls that genuinely require deep human judgment (complex B2B discovery, sensitive medical conversations). Use AI as triage and route to humans — don't expect it to close.
How to run your own number in 5 minutes
- Pull last 30 days of call logs from your phone provider. Count missed calls (no-answer + busy + abandoned voicemail).
- Pull last 90 days of revenue and divide by booked jobs to get average ticket.
- Estimate close rate honestly — if you recover a missed call, what's the chance it converts? Most owners overestimate by 10–20 points.
- Multiply:
missed × close × ticket × 12. That's your annual upside. - Subtract
AI cost × 12. That's your net.
If the net is positive and the upside is more than 5× the cost, it's a clear yes. If it's between 1–5×, run a 30-day pilot. If it's under 1×, your bottleneck isn't missed calls — it's something else.
Frequently asked questions
How accurate is the missed-call number from my phone provider?
Reasonably accurate but usually undercounts. Most carrier dashboards show no-answer + voicemail. They miss: calls that ring once and hang up (modern customers don't leave voicemails), calls during voicemail-full periods, calls during busy signals, and after-hours calls that customers don't bother with. The real missed-opportunity number is usually 1.5–2.5× the carrier dashboard number. The cleanest way to get the real number is to forward calls to a tracking service for 14 days, then count every inbound number that didn't connect to a live person. Most AI receptionist vendors offer a free trial that doubles as a missed-call audit — you can see exactly what was being missed.
What if my close rate is lower than the examples?
Lower the close-rate input and rerun. The formula still works — it just changes the break-even number. A 20% close rate on a $1,000 ticket at $300/month AI cost breaks even at 1.5 calls/month, which is still trivial for almost any service business. The close rate matters most at low ticket sizes. Below $200 average ticket with under 30% close rate, the math gets thin and you should pilot rather than commit. Also remember: close rate isn't fixed. Better qualification questions and faster follow-up SMS (both standard AI receptionist features) typically lift recovered-call close rates 5–15 percentage points [VERIFY] over the first 90 days as the script tunes.
Should I count voicemails as missed calls?
Yes, with a haircut. Of the customers who leave voicemails, only 30–50% actually become customers if you call back within an hour [VERIFY], and that drops fast as response time stretches. So the right modeling: count voicemails as missed calls, then apply a discount factor (multiply by 0.5 for a conservative number). Alternatively, count voicemails separately and assume the AI converts them at a higher rate than no-message hangups (because the caller already declared intent). For most service businesses, voicemails are a single-digit percentage of inbound — what really moves the ROI number is silent missed calls (ring-and-hang-up, busy-signal, after-hours).
Does the ROI math change if I already have a human receptionist?
Yes. You're not recovering missed calls from zero — you're recovering the ones the human can't take (overflow, after-hours, lunch breaks, sick days). The math becomes: how many calls per month does my receptionist miss, multiplied by close rate, ticket, and 12. For most single-receptionist businesses, that's 30–80 calls/month [VERIFY] depending on size and after-hours volume. The AI receptionist doesn't replace the human — it covers the gaps. Many businesses also use AI for first-touch qualification, freeing the receptionist to handle complex calls and walk-ins. The combined model often produces higher booking rates than either alone.
How long until I can prove ROI?
Most service businesses see clear positive ROI within 30 days, and the data is conclusive within 90. The fastest indicator is the booking-confirmation count — within the first week, you can count appointments booked by the AI that wouldn't have existed otherwise. Within 30 days, you can compare revenue from AI-sourced bookings against the AI's monthly cost. Within 90 days, you have enough data to compare close rates on AI-handled calls vs. live-answered calls and refine the script. Set a recurring monthly review: AI-sourced bookings × close rate × average ticket vs. AI cost. Most businesses stop running this calculation after month three because the answer is so consistently positive it stops being a question.
Ready to see what an AI receptionist looks like for your business? → Book a 20-minute demo