This is the third piece in the 2026 buyer's bar series. The first piece named the six standards every contractor AI answering service should meet. The second piece showed how to test them with five real calls. This piece does the math on the standard that breaks the most contractor budgets: pricing-model risk.
If your answering service charges per minute, your peak week is your most expensive week. That is the trap.
How per-minute billing works in practice
Every live answering service starts with a base monthly fee that covers a small number of included minutes, then bills per minute over the included quota. Ruby Receptionists publishes $319/month for 50 included monthly calls on its baseline and $599/month for 200 monthly calls on its larger plan (ruby.com/pricing, accessed 2026-05-15). Smith.ai publishes $293/month for 30 included monthly calls on Starter and $765/month for 200 monthly calls on Basic, with extras at $11.50 to $15 per call depending on plan and call duration (smith.ai/pricing, same date). PATLive publishes $235+/month for 75 minutes. AnswerConnect's published anchor is $325/month with per-minute overage. Nexa publishes $239/month with per-minute overage.
The per-minute meter is fine when call mix is even. The trap shows up in real contractor weeks where call volume and call duration both spike at the same time.
The contractor week that breaks the meter
Three scenarios you actually face every year:
Scenario 1: HVAC heat-wave Saturday. Outside temperature crosses 100F. Phone rings for ten hours straight: "AC died," "system blowing warm air," "no cool air." Average call duration is longer than a routine call because every caller wants confirmation on a same-day visit and rate quote. Call count: 35-50 in one day. Talk-time: 4-6 minutes per call vs the 2-3 minute baseline.
Scenario 2: Plumbing freeze night. First hard freeze in a cold-snap region. Burst pipes start at 11pm and continue through dawn. Every caller is panicked, has water flowing, needs shutoff guidance, needs a same-night dispatch confirmation. Call count: 20-30 in twelve hours. Talk-time: 5-8 minutes per call.
Scenario 3: Roofing hail event. A storm passes through your service area. The next morning, every homeowner who heard the noise wants an inspection appointment and a claim-process walkthrough. Call count: 50-100 in two days. Talk-time: 6-10 minutes per call because the caller wants insurance guidance.
These are not edge cases. These are the weeks contractors plan their year around. Three to five times a year, every regional service business has a week that defines their revenue.
The math at 200 monthly calls in a peak week
Same call mix, same talk-time. Three vendors compared with their published 2026 baseline plans (sources cited above):
| Vendor | Plan | Included | At 200 monthly calls | Annualized |
|---|---|---|---|---|
| OnCrew Pro | $149/mo | 400 calls included | $149 | $1,788/yr at this volume |
| OnCrew Starter + overage | $49 + 100 × $0.99 | 100 calls included | $148 | $1,776/yr (avg) |
| Smith.ai Basic | $765/mo | 200 monthly calls included | $765 (no overage hit) | $9,180/yr |
| Smith.ai Starter | $293/mo | 30 calls | $293 + 170 × ~$13 = $2,503 | $30,036/yr equivalent |
| Ruby (smaller plan) | $319/mo | 50 calls baseline | $319 + 150 × per-min = $1,400+ | $16,800+/yr |
| PATLive | $235+/mo | 75 minutes | $235 + ~400 metered min = $1,200+ | $14,400+/yr |
OnCrew Pro is the only line that does not change. The included call quota is 400; 200 monthly calls fits inside the plan. The price is $149 whether the calls happen evenly across the month or all on one heat-wave Saturday.
That is the structural difference. Flat per-call pricing decouples your monthly cost from your peak week. Per-minute pricing couples them together, in the worst direction for your cash flow.
Why this is harder to spot than it should be
The pricing pages of live answering services do not display the math above. They show the base monthly fee and the per-minute rate, in two different sections. A buyer evaluating the service does the math on a normal-volume month and decides "this seems OK." The first time the meter actually bites is the first peak week, and at that point the line item is already on the invoice.
Three honest disclosures every live answering service should make on the pricing page but typically does not:
- The actual per-minute rate after the included minutes are used. Some publish this; some only on a quote.
- Holiday surcharges. Several services charge extra on federal holidays. Contractors should know this before signing because freeze nights and heat waves often land on weekends and holidays.
- The "per call" or "per minute" billing definition. Some services bill per call regardless of duration; others bill the talk-time per call; others bill only the talked minutes, rounded up to the next minute. A 30-second voicemail captured by a live agent might cost a full minute on one service and a full call on another. The arithmetic changes by 2-3x depending on how the billing rule applies.
We publish OnCrew pricing in one place (pricing): $49, $149, or $349 per month for 100, 400, or 1000 included calls, $0.99 per call on overage, no per-minute meter, no holiday surcharge. The math has no surprises by design.
When per-minute pricing is still the right choice
This is not an argument against live answering services in every case. Per-minute pricing is the right choice when:
- Your peak week is your normal week (call mix is evenly distributed all year).
- A warm live human voice on every call is non-negotiable for your brand (high-end concierge, luxury, medical).
- Your call duration is consistently short (within 90 seconds is the sweet spot for per-minute billing).
- You want outbound calling from the receptionist on your behalf (this is a paid skill that flat-rate AI does not offer).
For most contractors fielding 100-500 calls per month with a mix of routine and emergency intake, the peak-week math runs against per-minute pricing within 3-6 months of forwarding.
How to read a competitor's pricing page in 60 seconds
Open the published pricing page. Find these three numbers:
- Base monthly fee.
- Number of "included" calls or minutes.
- Per-extra rate (per call OR per minute, note which).
Then ask: at my peak-week call volume, with my typical talk-time, how many extra calls or minutes will I cross? Multiply. Add to the base. Compare against OnCrew's flat plan at the same volume tier.
The bar should be the same for every service that wants to answer your business line. Pricing transparency is the first standard from the 2026 buyer's bar, the per-minute trap is what happens when that standard is hidden in fine print instead of headline copy.
If you would rather skip the math and run the actual five test calls from the 5-call audit, the free 60-question intake checklist gives you the trade-specific questions to evaluate each vendor on the call itself, not just the pricing page.
Sources
- Smith.ai pricing page, smith.ai/pricing, accessed 2026-05-15.
- Ruby Receptionists pricing page, ruby.com/pricing, accessed 2026-05-15.
- PATLive pricing page, patlive.com/pricing, accessed 2026-05-15.
- AnswerConnect pricing page, answerconnect.com/pricing, accessed 2026-05-15.
- Nexa pricing page, nexa.com/pricing, accessed 2026-05-15.
- OnCrew pricing source of truth: src/lib/billing/plans.ts in the public repo.