Ruby pricing (2026): the per-minute math

Ruby's current plans, verified (June 2026)

Ruby is one of the oldest names in live answering — real, US-based receptionists who pick up your line in your company's name. Its pricing is built on one unit: the receptionist minute. You buy a monthly block of minutes, and everything the receptionist does on your behalf draws the block down.

Here are the published virtual receptionist plans on ruby.com as of June 2026. Ruby states there are no fees for activation, onboarding, setup, or customization; the plans differ by the size of the minute block.

• 50 receptionist minutes — $250/month, which works out to $5.00 per minute (as of June 2026)
• 100 minutes — $395/month ($3.95 per minute)
• 200 minutes — $720/month ($3.60 per minute); the plan Ruby flags as its most popular
• 500 minutes — $1,725/month ($3.45 per minute)
• Live website chat is a separate product ($143–$520/month for 10–50 chats, as of June 2026); bundling chat with receptionist service takes 20% off the chat price
• Unused minutes do not roll over, and going over your block triggers a per-minute overage billed the following cycle

What a receptionist minute actually is

This is the part most buyers skim past, and it is where the real cost lives. Ruby bills receptionist time, not call time — but receptionist time is broader than the conversation itself.

Per Ruby's own billing documentation, the clock starts the moment a receptionist answers and includes time your caller spends on hold. It also includes after-call work: the time the receptionist spends completing your message after the caller has hung up. On outbound call assists, the follow-up email they send you is on the clock too. The clock stops when the call is transferred to you or to voicemail.

Then there is rounding. Ruby's current website FAQ states minutes are billed in 60-second increments, rounded up to the nearest 60-second mark — its example is that a 10-second call bills as 60 seconds. (Ruby's older downloadable billing FAQ described 30-second rounding; the current published rule is the stricter one.) In practice, a 40-second wrong number or a solicitor who gets past the greeting costs you a full minute at $3.45–$5.00.

Overage rates are not published on the pricing page. The sample invoice in Ruby's downloadable billing FAQ shows overage minutes at $3.60 each on a 500-minute plan — though that invoice is a few years old, so treat the figure as indicative rather than current — and third-party reviews report roughly $3.30 to $5.40 per minute depending on tier. If you go over your block, you are not bumped to the next plan automatically — you simply pay the overage.

The math: what $720 for 200 minutes buys an HVAC company

Work it forward with honest assumptions. A typical service-business call handled by an answering service — greeting, a few intake questions, maybe a brief hold, then the message write-up, with round-up on top — runs about three billed minutes. Some calls are two, a thorough intake or a scheduling call is four or five.

At three billed minutes per call, the $720 plan's 200 minutes cover roughly 65–70 handled calls a month. That is about three calls per business day, at an effective cost of around $10.80 per call. On the 50-minute plan, the same call costs about $15.

Now apply HVAC seasonality. A two-truck shop averaging eight calls a day in shoulder season takes around 240 calls a month — roughly 720 receptionist minutes. That already overruns the top published 500-minute plan; at the $3.60 overage shown on Ruby's sample invoice, the month lands near $2,500. In a July heat wave at fifteen calls a day, you are at about 1,350 minutes and roughly $4,800 for the month.

And because unused minutes do not roll over, a quiet April does not bank anything for July. The structure means your bill peaks in exactly the months your phone does.

What Ruby genuinely does well — and who should pick it

None of this math makes Ruby a bad service. It is arguably the best at what it actually sells: a warm, professional human voice answering in your name. Ruby has been doing this since 2003 — two decades of training receptionists in tone, empathy, and judgment. A live person can read a distressed caller, handle an unusual situation gracefully, and represent a high-end brand in a way no script can. Answering runs 24/7/365, live transfer works well, and bilingual English/Spanish answering is available on weekdays (5am–6pm Pacific).

Ruby is a strong fit for law firms, financial practices, and professional services where callers expect a person, call volume is modest and mostly business-hours, and the revenue per client makes $3.60–$5.00 a minute immaterial. If you take fifteen meaningful calls a week and each retained client is worth five figures, Ruby's model is fine — arguably ideal.

The strain shows up when the model meets trades volume. Ruby receptionists can book appointments if you connect a scheduler, but every minute of booking is metered, so many trades accounts settle into message-taking and relay — which means you still return the call to win the job. After-hours answering is available, but a 2 a.m. no-heat call bills the same minutes as a 2 p.m. one. And the per-minute meter means cost scales linearly with exactly the demand spikes a trades business lives on.

The alternative model: AI agents with flat or near-flat pricing

AI answering inverted the cost structure. Because software answers the call, vendors can price flat or per-caller instead of per-minute — and a ten-minute booking call costs the vendor little more than a two-minute one. The trade-off is that you give up human judgment on the line; the better systems offset that by booking jobs directly on your calendar, texting callers back, and escalating edge cases to you. Verified pricing as of June 2026:

• Ruby (for reference) — $250–$1,725/month for 50–500 receptionist minutes (as of June 2026); real human receptionists, two decades of polish, live transfer; best for professional services with modest, business-hours call volume
• Smith.ai AI Receptionist — $95/$270/$800 per month for roughly 2/5/15 calls a day, $2.40 per call overage, $3 per call to escalate to a live agent (as of June 2026); strong intake workflows and a credible human-escalation path; best for firms that want AI economics with humans on standby
• Goodcall — $79/$129/$249 per month per agent with no per-minute charges, billed on unique callers (100/250/500 included, then $0.50 per extra unique caller) (as of June 2026); simple setup, no minute meter at all; best for small shops that want a basic AI answering layer cheaply
• Rosie — $49/$149/$299 per month for 250/1,000/2,000 minutes, with calendar booking from the $149 tier (as of June 2026); inexpensive, fast to try, bilingual; best for very-low-volume businesses testing AI answering for the first time
• Power2Network — $1,000 one-time build plus $499/month flat, unlimited answering, cancel any month, carrier/usage costs passed through at cost (as of June 2026); an AI workforce, not just a receptionist: every call answered 24/7, jobs booked on your calendar, missed-call text-back, follow-up and reviews, managed by a named human. Candidly: it costs more than the DIY tools above, and if you take a handful of calls a week, Rosie or Goodcall is the smarter buy. P2N is built for businesses where a missed call is a lost job worth hundreds or thousands — and where July should not cost ten times what April does

The bottom line for a trades business

Ruby's per-minute model rewards low, steady, business-hours volume and penalizes high, spiky, around-the-clock volume. That is the opposite shape of an HVAC, plumbing, or pool company's phone traffic. If your callers are weighing six figures and expect a human voice, pay Ruby happily. If your phone is how jobs arrive — and it rings hardest at night, on weekends, and in season — a flat-rate system that answers everything and books the job tends to win the math.

For what that looks like over time rather than in a demo: Service One, an HVAC company in Hampstead, NH, has run Power2Network's missed-call recovery continuously for more than two years — one named client's documented outcome, not a typical result. The write-up is in our results section. Run your own numbers: count last July's calls, multiply by three minutes, and price that month on any per-minute plan before you sign one.