Skip to content
The AI Agent ReportFind My AI Agent Path

Pricing guide · 12 vendors · Documentation review

AI Receptionist Cost in 2026: What 12 Vendors Charge

Last reviewed: Editor: Jordan M. ReyesEvidence level: Documentation review

Pricing verified on vendor pricing pages on . No hands-on call testing is claimed on this page. Some vendor links may earn a commission — pricing, rankings, and critical coverage are not adjusted for affiliate compensation. Full disclosure.


The short answer

Most U.S. small businesses pay $79–$249 per month for an AI receptionist that books appointments, qualifies leads, and integrates with a CRM or calendar. The cheapest plans start at $25–$29/month but cap calls or minutes hard. Developer voice-agent platforms advertise $0.05–$0.09 per minute but real all-in cost lands between $0.13 and $0.40 per minute once you add speech-to-text, the language model, voice generation, and telephony. Hybrid AI-plus-human services run $250–$800/month.

Your real bill depends less on the sticker price than on three things you control: your monthly call volume, your average call length, and the billing model the vendor uses to charge you. Get those wrong and a “$99/month” plan becomes $400.

TierMonthly costWhat you getBest for
Budget AI$25–$79/moBasic call answering, limited minutes, simple FAQs, may lack integrationsTesting the category, very low call volume (under 30/mo)
SMB sweet spot$79–$249/moAppointment booking, CRM sync, customizable scripts, decent voicesMost small businesses (30–150 calls/mo)
Premium AI / Hybrid$250–$800/moHigher call allowances, hybrid AI + human, advanced routing, white-glove onboardingProfessional services, multi-location SMBs
Developer per-minute$0.07–$0.40/min all-inMaximum control, custom workflows, build your own receptionist behaviorTechnical teams, agencies, high-volume operators
Enterprise / ComplianceCustom (typically $1,000+/mo)HIPAA BAA, SOC 2 review, SLAs, dedicated infrastructure, SSORegulated industries, multi-location healthcare, large legal firms

Direct budget recommendation

Most operators reading this page should budget $100–$300 per monthfor a real AI receptionist pilot. If your workflow needs human backup, complex CRM logic, multiple locations, HIPAA review, or long consultative calls, model the total cost before assuming AI will be cheaper. Sometimes it won’t be.

If you’re missing after-hours calls at a small home services or professional practice, the entry tier ($79–$129/mo) almost always pays for itself by catching one to two extra jobs per month.

Five-second decision table

Your situationLikely best modelVendors to inspect firstWhat to watch for
Simple front desk, predictable volumeFlat monthly planGoodcall, Frontdesk, DialzaraUnique-customer caps, included minutes
Need human backup on important callsPer-call with human escalationSmith.ai, RubyPer-call overage rates
Technical team, custom workflowsPer-minute developer platformRetell, Synthflow, Vapi, BlandReal all-in cost, not platform fee
Regulated workflow (healthcare, legal, financial)Compliance-ready plan or enterpriseBland, Smith.ai, SynthflowBAA, SOC 2, data retention, counsel review
Just testing after-hours coverageLow-cost pilot, simple workflow firstFrontdesk, Goodcall, SonaDon't test on emergencies first

The AI receptionist true-cost calculator

Bottom-line answerA real cost estimate requires four inputs: monthly call volume, average call length, percentage requiring transfer or human handoff, and compliance needs. Without those, “AI receptionist cost” is just a sticker price.

Run your numbers through the model below. The output rows show your estimated monthly bill across all five pricing models and identify the model that fits your usage pattern.

Your inputs

Monthly inbound call volume

10–1,000 calls

Average connected call length

1–10 minutes

Calls requiring booking or transfer

0–100%

Need human fallback on complex calls

Yes / No

Regulated workflow requiring BAA

None / Healthcare / Legal / Financial

Need CRM and calendar integration

Yes / No

Estimated monthly cost by model — 3-minute average call

Billing model30 calls / 90 min100 calls / 300 min300 calls / 900 minBest when…
Flat monthly (Frontdesk, Goodcall, Dialzara)$25–$99$99–$129$149–$249Volume is predictable and you want a fixed line item
Per-unique-customer (Goodcall)$79 (≤100 unique callers)$79–$129$249Repeat callers with long calls; unique-customer cap not hit
Per-call AI (Smith.ai AI Receptionist)$95 (Starter, in plan)$215–$270$630–$800Calls average 2+ min; volume is steady
Per-minute managed (Ruby, Dialzara overage)$250 (Ruby 50-min plan)$395–$720$1,210+Call length varies wildly; want predictable per-minute rate
Developer platform all-in (Retell, Bland, Vapi)$12–$36 at $0.13–$0.40/min$39–$120$117–$360Technical team; high volume; want to swap LLMs per use case

“Avoid this model if…” warnings

Flat monthly:Your call volume spikes unpredictably — overage charges can easily double the base price.
Per-unique-customer:You get high volumes of one-time callers (cold leads, walk-in requests) — the cap hits faster than expected.
Per-call AI:Your calls are very short (under 2 min avg) — per-minute wins below that threshold. Also: spam calls may count.
Per-minute managed:Your intake calls run long (6+ min) — per-call beats per-minute above the 2-minute breakeven.
Developer platform:You have no engineering capacity — the platform fee is not the full cost; you must model STT, LLM, TTS, and telephony separately.

Calculator assumptions (so you can verify the math)

Per-minute managed estimates use a $0.30/min midpoint

Per-call estimates use a $1.90/call midpoint with $2.40 overage above plan cap

Developer-platform estimates use a $0.13/min all-in floor and $0.25/min midpoint

Flat-rate estimates use vendor-published included minutes and current overage rates

Estimates are directional and were last refreshed on May 20, 2026


Verified AI receptionist pricing: 12 vendors, same columns

Every vendor below is priced against the same columns: billing model, starting price, included usage, overage rate, setup fee, transfer fee, hidden-cost flags, evidence level, and source. Same columns for everyone — no vendor gets a column the others don’t get. Pricing verified on each vendor’s own pricing page on .

VendorBilling modelStarting priceIncluded usageOverage rateSetup feeTransfer / escalationHidden-cost flagsEvidenceSource
Smith.ai AI ReceptionistPer-call subscription; AI-first with optional human escalation$95/mo Starter~2 calls/day on Starter; per-call equiv. ~$1.90 (Starter), $1.80 ($270), $1.60 ($800)$2.40/call$0 (self-service)On-demand human handoff billed as add-onDone-for-you annual plans from $500/mo; integration depth varies by tierDoc reviewsmith.ai/pricing/ai-receptionist
Smith.ai Virtual ReceptionistsHuman-staffedPer-call subscription; human-first$2,100/mo (300-call plan displayed)300 calls/mo at displayed tier$8.50/call overageNone standardBuilt into per-callLower tiers and enterprise custom pricing exist; verify current tier directlyDoc reviewsmith.ai/pricing/receptionists
GoodcallPer-agent subscription with unique-customer thresholds$79/mo Starter100 unique customers/mo, unlimited minutes and tokens$0.50/extra unique customer$0BundledPer-agent multiplier on team accounts; re-confirm current tier names on live pricing pageDoc reviewgoodcall.com/pricing
Frontdesk / My AI Front DeskSubscription + credit pool$99/mo ($79/mo billed annually)200 voice minutes, 100 chatbot conversations, 400 SMS, 1,000 monthly overage credits; Zapier, verified outbound number, 3 notification recipientsVoice: 25 credits/min; SMS: 4 credits/msg; chatbot: 5 credits/conversation; credits at $0.01 each$0Counts against creditsFree plan exists with limits; "$79/mo" requires annual prepay ($948 up front)Doc reviewmyaifrontdesk.com
Retell AIPay-as-you-go per minute$0/mo + $0.07/min and upFree starting credits; 20 concurrent calls included; PII redaction, opt-out recording$0.07–$0.31/min depending on voice and LLM stack$0Bundled at per-minute rateConfirm which components are included in your selected configurationDoc reviewretellai.com/pricing
SynthflowPay-as-you-go per minute$0/mo to create account; usage charges apply when calls are sentNone default$0.15–$0.24/min depending on LLM and telephony selections$0BundledConcurrency limits, model upgrade costs, enterprise add-onsDoc reviewsynthflow.ai/pricing
VapiPlatform fee + provider pass-through (developer platform)$0.05/min platform feeNone — STT, LLM, TTS, telephony each billed by their providers separately$0.05/min platform + provider rates; real all-in: $0.13–$0.40/min (modeled estimate from third-party synthesis)$0Separate per-minutePlatform fee is not the total cost. Must model STT + LLM + TTS + telephony separatelyDoc review + modeled estimatevapi.ai/pricing + third-party synthesis (Klariqo, pxlpeak, Telnyx)
Bland AIPlan fee + per-minute usageStart: $0/mo at $0.14/min · Build: $299/mo at $0.12/min · Scale: $499/mo at $0.11/minNone by defaultPlan rates apply per minute$0Start: $0.05/transfer min (Bland telephony); Build: $0.04/transfer min; free on BYOT (Bring Your Own Twilio)$0.015 min charge on outbound attempts under 10 sec; $0.02/SMS; old "$0.09/min" rate was deprecated late 2025 — older comparisons are staleDoc reviewdocs.bland.ai/platform/billing + bland.ai/pricing
Ruby ReceptionistsHuman-staffedPer-minute bundled, human-staffed$250/mo (50 min)50 min ($250) · 100 min ($395) · 200 min ($720)$5.40/min (50-min plan) · $4.50/min (100-min) · $4.40/min (200-min)None standardBuilt into per-minutePremium bilingual and after-hours rates may apply; verify current rounding rules in termsDoc reviewrubyhelpcenter.helpjuice.com
DialzaraSubscription with bundled minutesNeeds direct verificationVerify before signingVerify before signingVerify before signingVerify before signing7-day free trial confirmed; pricing tiers needed from current pricing page. Do not rely on this row for budget planning without direct verification.Needs primary verificationdialzara.com
RingCentral AI ReceptionistAdd-on to RingEX seats$39/mo (GA announcement); $59/mo cited elsewhere — verify current pricing page100 min/mo per GA announcementVerify current rateTied to existing RingEX subscriptionBundledRequires RingEX base subscription; pricing references are inconsistent across RingCentral pages — verify before signingDoc review — conflict notedringcentral.com (verify standalone vs RingEX-bundled)
Sona by QuoCredit-based subscription$25+ (1,000 credits/plan ≈ 10 calls/mo)Calls under 15 seconds don’t countOverages disabled by default unless enabled$0BundledAdditional phone numbers extra; verify current credit allowances directlyDoc reviewsupport.quo.com/sona-pricing

What we actually verified on May 20, 2026

We did verify: Public pricing pages, plan tier structures, billing units, stated included-usage limits, published overage rates, publicly stated setup fees, publicly disclosed transfer fees, and vendor-published integration claims.

We did not verify: Real call accuracy, booking accuracy, latency, hallucination rate, escalation success rates, support responsiveness, actual BAA negotiation process, or real-world caller satisfaction. We do not claim hands-on call testing on this page.

See full methodology →


The five AI receptionist pricing models

Bottom-line answerAI receptionist providers use five distinct pricing models. Picking the wrong one for your call profile can double or triple your bill before you notice. The right model depends on whether your bottleneck is call count, call length, unique customers, workflow complexity, or compliance.
1

Flat monthly subscription

Frontdesk, Sona, Dialzara (verify current plans)

The standard SMB model.You pay a fixed monthly fee that includes a defined number of calls, minutes, or features. Predictable, easy to budget, but “unlimited” claims sometimes have fair-use caps and included usage can be tight.

Cost range

$25–$500/month

Best for

Operators with 30–300 calls/month and predictable volume; office managers who need a fixed line item

Watch-outs

Annual-only lock-ins; fair-use clauses on “unlimited” plans; add-on fees for integrations on lower tiers

2

Per-unique-customer pricing

Goodcall (only major vendor using this as primary unit)

You pay based on how many distinct callersthe AI talks to in a month, not how long it talks to them. Unique-customer pricing decouples call length from cost — a 20-minute call doesn’t punish you if the customer already counts as a unique caller this month.

How it works

Monthly subscription + unique-customer cap. Repeat callers don’t double-count. Goodcall Starter: 100 unique customers, unlimited minutes.

Best for

Smaller customer base making longer calls: counseling practices, B2B firms, returning patient workflows

Watch-outs

High volume of one-time callers (cold leads, walk-ins) hits the cap fast. Multi-team setups multiply per-agent cost.

3

Per-call pricing (often with human escalation)

Smith.ai AI Receptionist, Smith.ai Virtual Receptionists

You pay per call answered, regardless of length. Smith.ai’s self-service AI Receptionist runs $95/mo, $270/mo, and $800/mo, with per-call equivalents of roughly $1.90, $1.80, and $1.60 and an overage rate of $2.40/call. Done-for-you annual plans start at $500/mo. Their human-staffed Virtual Receptionists currently show a 300-call tier at $2,100/mo with $8.50/call overage.

Break-even vs per-minute

Per-call wins above ~2-minute average call length; per-minute wins below that

Best for

Professional services (law, accounting, real estate) where calls average 2+ minutes and human backup matters

Watch-outs

30-second hangup costs same as 5-minute booking. Spam calls may count against quota. On-demand human handoff is usually a separate per-call add-on.

4

Per-minute managed pricing

Ruby Receptionists, Dialzara (verify), most live answering services

You pay for actual talk time at a per-minute rate, often with a plan fee that lowers the per-minute cost. Live answering services from Ruby Receptionists run $250–$720/mo for 50–200 included minutes, with overages of $4.40–$5.40/min depending on plan. Per-minute can be cheap if calls are short or expensive if callers linger.

Cost range

$0.10–$0.50/min advertised range; Ruby overages $4.40–$5.40/min

Best for

Operators with highly variable call volume; businesses where call length varies wildly

Watch-outs

Voicemails often count as billed minutes. Failed calls can carry minimum charges. Spam calls usually count.

5

Developer voice-agent platform

Vapi, Retell, Bland, Synthflow

You pay a small platform fee plus pass-through costs for every component of the AI stack — speech-to-text, the language model, voice generation, and telephony. This is the model that looks cheapest on paper and is most often misunderstood.

The Vapi math, shown:

$0.05/min (platform) + ~$0.005/min (speech-to-text) + $0.02–$0.20/min (LLM, varies by model) + ~$0.04/min (ElevenLabs TTS) + ~$0.01/min (Twilio telephony) = $0.13–$0.40/min all-in (modeled estimate from third-party operator data, not vendor-published as single figure).

Retell:

Bundles platform + LLM + voice at $0.07–$0.31/min depending on stack selected.

Bland:

Start $0.14/min · Build $299/mo + $0.12/min · Scale $499/mo + $0.11/min. Old $0.09/min universal rate was deprecated late 2025 — older comparisons are stale.

Best for

Technical teams with engineering capacity; agencies building custom voice agents; high-volume operators where per-minute economics beat any subscription

Also good for

Anyone who wants to swap LLMs per use case or run multi-agent flows

Critical watch-out

The platform fee is not the cost. Vapi’s “$0.05/min” is real but you’ll pay several times that all-in. Failed calls carry minimum charges on Bland. Voicemails billed at standard rate.


What the sticker price hides

Bottom-line answerThe advertised price is not the bill. Setup fees, transfer-time charges, failed-call minimums, SMS line items, integration tier-ups, and compliance add-ons can add 20–40% to your real monthly cost.
Most AI receptionist pricing pages are designed to make the sticker look small. The only honest comparison is one done with the same columns across every vendor — which is what we built in the verified matrix above. One thing we’ll say plainly that most pricing guides won’t: the AI receptionist category is not “AI is always cheap.” It’s “AI is cheaper than a human for the call types AI handles well, and roughly the same price as a live answering service for everything else.”

Setup and onboarding fees ($0–$5,000)

Most modern AI receptionists charge $0 in setup fees. Custom integrations, white-glove onboarding, or enterprise deployments can range from $99 to $4,999+ depending on the vendor. Always confirm setup posture directly with the current pricing page or sales rep before signing.

Transfer / live handoff fees ($0–$3 per call or per minute)

When the AI hands a call to a human, that's often billed separately. Bland charges $0.05/transfer min (Start) and $0.04/transfer min (Build) on its own telephony, with transfers free on Bring Your Own Twilio. Smith.ai's on-demand live human handoff is a separate add-on. Most flat-rate plans bundle internal transfers but charge for external escalation.

Failed-call and outbound minimums ($0.015–$0.50 per attempt)

Bland publishes a $0.015 minimum charge for outbound attempts that fail or end in under 10 seconds. This isn't unusual — it covers the carrier handshake — but it surprises new buyers running cold outbound campaigns at low connect rates.

SMS and email line items ($0.01–$0.05 per message)

AI receptionists that send appointment reminders, follow-ups, or text links charge separately per message. Bland is $0.02/SMS. Frontdesk consumes credits from the same pool that voice minutes draw from (4 credits per SMS). Budget for it if SMS is part of your workflow.

Integration / connector fees

Native CRM integrations are typically free on mid and high tiers. "Via Zapier" means you're paying Zapier separately — usually $20–$50/month for the connector volume you'll need. Some vendors gate calendar booking, CRM sync, or SIP trunking behind higher plan tiers.

Compliance and BAA fees

Some third-party comparisons report a Vapi HIPAA add-on around $1,000/month — treat that as third-party-reported and verify with Vapi directly before publishing as a current compliance cost. Most managed AI receptionists bundle BAA into higher plans without separate fees, but the BAA process can take days to weeks depending on the vendor's legal team.

Annual vs. monthly lock-ins

The advertised "$79/mo" sometimes requires annual prepay ($948 up front). My AI Front Desk's $79 Starter rate is "billed annually." Read the billing-frequency footnote before signing.

Spam-call quota erosion

When spam calls count against your monthly quota, your effective cost per real call rises. Vendors handle spam differently — some block known robocalls automatically, some count every call against your quota by default, some require Caller ID forwarding to be configured.

Concurrency limits

If your business gets a burst of simultaneous calls — a marketing campaign drops, a news mention drives volume, a season starts — concurrency limits can drop calls or queue them. Higher plans buy higher concurrency. Retell publishes 20 included concurrent calls on pay-as-you-go. Bland and Synthflow list concurrency by plan.


AI receptionist vs human receptionist vs live answering service

Bottom-line answerA full-time human receptionist costs $3,000–$5,000 per month loaded. A live answering service runs $250–$1,400+ per month. An AI receptionist that handles the same job for most small businesses runs $79–$249 per month— often materially cheaper on routine call coverage, with the savings shrinking when calls require human backup, long talk time, or compliance controls.

Full-time human receptionist: $3,000–$5,000/month loaded

BLS lists the 2024 median receptionist wage at $37,230/year ($17.90/hour) before employer taxes, benefits, paid time off, training, and turnover. Using a 25–35% employer-overhead assumption, that models to roughly $46,500–$50,300/year — or $3,875–$4,200/monthat the base. Add backup, training, and software access and many practices land at $4,500–$5,000/month.

You get

Judgment. Empathy. A human who handles a crying customer, a confused elderly caller, a complex insurance question. Someone who learns your business over months.

You don’t get

24/7 coverage without multiple shifts (doubles/triples cost). Sick day coverage. Handling several simultaneous calls. Predictable cost during volume spikes.

Live answering service: $250–$1,400+/month

Real humans answer your calls, sharing attention with dozens of other businesses. Ruby starts at $250/mo for 50 minutes. Smith.ai Virtual currently shows a 300-call plan at $2,100/mo. Overages: $4.40–$5.40/min at Ruby; $8.50/call at Smith.ai.

You get

Human empathy without the full-time salary. After-hours coverage. Reasonable handling of complex intake.

You don’t get

Same-business familiarity. Predictable cost during busy months. Instant answer times (queue waits common).

AI receptionist: $25–$800/month managed

You get

24/7 availability. Higher concurrency than a single receptionist. Predictable cost on flat-rate plans. Direct calendar booking, CRM sync, and follow-up automation.

You don’t get

Real empathy on a hard call. Reliable judgment on edge cases. Crisis-level conversation handling. (Improved a lot in 2025–2026 but not perfect.)

Side-by-side comparison

DimensionFull-time humanLive answering serviceAI receptionistHybrid AI + human
Cost shape$3,875–$5,000/mo loaded$250–$2,100+/mo$79–$800/mo or per-minute$250–$1,000+/mo
Coverage hours40/week unless staffedPlan-dependent (often 24/7)24/724/7 with human escalation
Concurrent calls1Limited by their staffHigher than 1 (plan limits)Higher than 1, queued human
Empathy / nuanceHighMedium-highLow-medium (improving)High (when escalated)
Compliance handlingManual, error-proneTrained in-houseVendor-dependent (BAA, SOC 2)Vendor-dependent
Scaling costLinear (hire more)Linear (overage)Flat or sublinearMixed
Sick days / vacationReal problemBuilt-in (their problem)NoneNone

Break-even formula

Break-even missed calls/month = Monthly AI receptionist cost ÷ (Avg gross profit per converted call × Close rate)

$99/mo plan

$99 ÷ ($400 × 0.25)

~1 incremental conversion/month

Entry tier pays for itself quickly if calls have real value

$249/mo plan

$249 ÷ ($400 × 0.25)

~2.5 incremental conversions/month

Most SMBs hit break-even in the first week of the month

$800/mo plan

$800 ÷ ($400 × 0.25)

~8 incremental conversions/month

High-volume practices with structured workflow

Assumptions: $400 avg gross profit per converted call, 25% close rate. Adjust to your actual numbers before using this to justify a budget.


What should you budget at 30, 100, or 300 calls per month?

The three most common operator profiles. Find yourself in the answer.

Profile 1

Solo professional — 30 calls per month

Solo attorney · single-provider dental practice · independent consultant · freelance contractor · single-location salon

Likely bill

$25–$99/mo

Best billing model

Flat monthly entry tier. Don’t pay for capacity you won’t use.

Call profile

20–40 calls/mo · 2–4 min avg · mostly inbound · low compliance burden

Vendor shortlist:

Sona by Quo ($25+ credit-based) — calls under 15 seconds don’t count, predictable per-call style billing

Frontdesk free or Business ($0 or $79/mo annual / $99/mo monthly) — test before you commit

Goodcall Starter ($79/mo) — if you have repeat customers and longer calls

⚠ Call caps on cheap tiers fill up fast during seasonal spikes. If your calls average 5+ minutes (consultative practice), check minute-based plans — per-call can be cheaper than per-minute for you.

Profile 2

Small business — 100 calls per month

Multi-provider medical practice · small law firm · multi-location home services · medspa with steady booking traffic · mid-sized real estate brokerage

Likely bill

$99–$249/mo

Best billing model

Flat monthly mid-tier or per-call subscription, depending on call length

Call profile

80–150 calls/mo · 3–5 min avg · mix of booking, FAQs, lead qualification · needs CRM + calendar

Vendor shortlist:

Frontdesk Business ($99/mo) — 200 voice minutes, full integration suite

Goodcall Growth ($129/mo) — wider unique-customer headroom, unlimited minutes

Smith.ai AI Receptionist Starter ($95/mo) — if you want optional human handoff on important calls

RingCentral AI Receptionist add-on — verify current pricing if you’re already on RingEX

⚠ Per-agent multipliers on team accounts. Confirm CRM integration is included in your plan tier, not gated behind a higher tier. If calls regularly exceed 5 minutes, model per-call vs per-minute all-in.

Profile 3

High-volume practice — 300+ calls per month

Multi-location dental group · busy HVAC/plumbing contractor · mid-market legal firm · regional medspa chain · growing veterinary practice

Likely bill

$250–$800/mo managed

or $0.09–$0.15/min dev-stack (~$300–$900/mo at this volume)

Best billing model

Premium flat-rate, per-call upper tier, or developer per-minute (if you have engineering resources)

Call profile

250–500+ calls/mo · 3–6 min avg · heavy booking, complex routing, often HIPAA-sensitive

Vendor shortlist:

Goodcall Scale ($249/mo) — wider unique-customer headroom, unlimited minutes

Smith.ai AI Receptionist Pro or high-volume tier ($270–$800/mo) — human handoff available

Bland Build or Scale ($299–$499/mo + per-minute) — for technical teams building custom flows

Retell at scale — $0.07–$0.31/min depending on stack; roughly $300–$500/mo at 3,000–5,000 minutes

Synthflow Enterprise — for multi-location with configurable workflows

⚠ Concurrency limits on cheaper plans. Transfer fees stack at volume. Lock down BAA in writing before processing PHI. Some vendors quote enterprise pricing only after a sales call — budget extra time.

Vertical compliance overlay

Healthcare (HIPAA)

Require executed BAA before processing PHI. Treat any vendor-specific BAA claim as ‘needs vendor verification before PHI use’ unless you’ve reviewed primary documentation directly.

Legal intake

Per-call billing fits long intake calls. Conflict-of-interest screening is a feature, not a bullet. Smith.ai has the deepest legal-vertical integration depth.

Home services

Emergency-routing logic matters more than per-minute price. The AI must recognize “burst pipe,” “no heat,” “ASAP” as escalation triggers, not just routine appointment requests.

Outbound calling (any vertical)

TCPA and AI-voice disclosure rules apply. Get counsel review before any outbound AI campaign.


When is an AI receptionist actually worth it?

Bottom-line answerAn AI receptionist is worth it when it captures enough missed calls, saves enough staff time, or improves after-hours coverage enough to exceed the monthly cost. The strongest use cases are high-intent inbound calls where a fast answer matters and the workflow is structured.

Worth-it signals

You're missing calls after hours or during busy periods.
You get repetitive phone questions ("Are you open?", "How much does X cost?").
Your staff is interrupted by basic calls that could be handled by a script.
Your booking workflow is structured (defined services, providers, scheduling rules).
Your close rate from phone leads is meaningful — a captured lead is worth materially more than $200.
You need around-the-clock coverage but can't staff it.

Not-worth-it signals

Very low call volume (under 10/month).
No clear workflow yet — you don't know what the AI should say.
Complex emotional intake that demands empathy on every call.
High regulatory burden (controlled substance prescriptions, financial advice, complex insurance) without a vendor that can document the right compliance posture.
You cannot tolerate any AI mistakes — every failed call has six-figure downside.

Safe vs. unsafe first-test workflows

Safe pilot workflows

After-hours missed-call capture
"Are you open / what are your hours?" type FAQs
Appointment-request collection (not direct booking on first test)
Lead capture with handoff to a real callback
Confirmations and reminders (with reviewed consent language)
Simple routing ("press 1 for billing")

Unsafe first-test workflows

Emergency triage (medical, plumbing, locksmith)
Complex medical intake or symptom triage
Legal advice
Financial advice or coverage questions
Insurance verification (high error cost)
Refund disputes
Angry-customer escalation
Any regulated outbound campaign (TCPA exposure)

First 30-day pilot checklist

1Define one specific call type to test first (after-hours overflow is usually safest).
2Write the script in plain English — what the AI says, when it transfers, when it takes a message.
3Test five real call scenarios yourself before any customer hears it.
4Wire SMS notifications to the right person on your team for every captured lead.
5Review the first week of call transcripts daily. Catch the failure modes before they scale.
6Measure: calls captured, booking accuracy, escalation accuracy, customer complaints.
7Decide at day 30: expand workflows, swap vendors, or roll back.

The right pilot is the one where a wrong AI response costs you a single re-routed call, not a regulatory action or a six-figure judgment.


The 10 questions to ask every vendor before you sign

Bottom-line answerA good quote audit turns vague pricing into measurable risk. Copy these 10 questions and send them to any vendor sales rep before you sign anything. Vendors that answer them quickly and in writing are easier to work with. Vendors that dodge them are telling you something.
1

What exactly counts as a billable call, customer, or minute?

Is a 10-second hangup billable? Is voicemail billed at the standard rate?

2

Are spam calls, hangups, transferred calls, and failed outbound attempts billed?

Get specifics in writing.

3

What's included in the base plan before overage charges begin?

Get the exact minute, call, and credit allowances.

4

What's the overage rate, and how is overage calculated?

At the moment of overage, or end-of-month true-up?

5

Are human transfers, escalations, and SMS follow-ups included or billed separately?

6

Are CRM and calendar integrations included in my plan tier, or do they require an upgrade?

7

What customer data is stored, for how long, and where?

Is there a documented data retention and deletion policy?

8

Is a BAA, DPA, SOC 2 report, or other security documentation available — and what's the turnaround to get it executed?

9

What are the cancellation, refund, and annual-contract terms?

Is there a minimum commitment? Is the advertised price contingent on annual prepay?

10

What happens during volume spikes?

Does the AI drop calls, queue them, escalate them, or pass them through with higher overage charges?

If a sales rep gives you a verbal answer to any of these, get it in email before you sign. SaaS billing disputes are won and lost in writing.


What does compliance add to AI receptionist cost?

Bottom-line answerCompliance-sensitive workflows can add real cost through plan upgrades, BAA review, security documentation, data retention controls, opt-out handling, and counsel review. This page publishes software-buying research, not legal advice — verify your specific obligations with qualified counsel before deploying AI in regulated workflows.

TCPA and AI voice calls (federal)

The TCPA governs unsolicited calls and texts to consumers. The FCC has ruled that AI-generated voice calls fall under TCPA artificial-or-prerecorded-voice restrictions. Outbound marketing, reminder, and follow-up workflows should be reviewed with counsel for consent, identification, and opt-out requirements.

Inbound calls — calls a customer makes to your AI receptionist — generally raise a different TCPA risk profile than AI-initiated outbound calling. The bigger TCPA exposure is on the outbound side: AI-powered appointment reminders, follow-up calls, callback automation, and outbound sales sequences.

Consent revocation

The FCC has emphasized that consumers may revoke consent in a reasonable manner, and callers cannot be forced into only one revocation path. Practical implication: if a caller says “stop calling me” during an outbound call, your AI must register that as opt-out — even if your default opt-out flow expects “press 9” or “text STOP.” This is a feature question for every vendor you evaluate.

State AI disclosure laws

State AI disclosure laws are inconsistent and many are narrower than people assume. For example, California’s bot disclosure law (SB-1001) is focused on online bot interactions used to mislead consumers in specified purchase or voting contexts — it’s not a blanket rule for every phone call. Other states have proposed broader rules. The landscape changes quarterly — verify the current law in every state where you take calls with qualified counsel.

Best practice regardless of legal minimum: have your AI receptionist identify itself as an AI assistant near the start of the call. It builds trust and pre-empts most state disclosure concerns.

HIPAA, BAA, and regulated workflows

If your AI receptionist will handle PHI — appointment scheduling that includes diagnoses, intake forms, symptom triage, anything beyond “what are your hours” — you need an executed BAA (Business Associate Agreement) with the vendor before going live.

A vendor saying “we are HIPAA compliant” is not the same as having executed a BAA with you. HIPAA compliance is a posture; the BAA is the contract that creates legal obligation. Don’t process PHI until the BAA is signed.

For any vendor-specific BAA or HIPAA claim, verify the current scope and process directly from the vendor’s own security or trust documentation, not from third-party comparison pages — including this one.

Practical compliance cost impact

TCPA outbound counsel review:Budget time for a one-time legal review before any outbound campaign. Actual cost varies by counsel and scope.
BAA execution:Usually $0 incremental fee from the vendor (the BAA itself is contractual), but vendor plan upgrade may be required to access BAA-eligible tiers.
SOC 2 documentation requests:SOC 2 reports are usually gated behind vendor security review or enterprise sales motion. Confirm availability directly with the vendor.
Data retention configuration:May require higher-tier or enterprise configuration. Verify vendor-specific pricing.

Budget compliance costs as a separate line item, not a feature of the AI receptionist plan.


Is there a free AI receptionist?

Bottom-line answerPermanent free AI receptionist tiers exist but are capped tightly — useful for testing, not for production. Most major vendors offer free credits or limited trials with full feature access, which is the better way to evaluate.

Free trial periods

Goodcall (verify current trial length) · Synthflow (free account; usage charges when calls are sent) · Dialzara (7-day free trial confirmed)

Free starter tier

Frontdesk / My AI Front Desk offers a free plan with limited usage. Verify current free-tier allowance directly before relying on it.

Pay-as-you-go with free credits

Retell starts at $0 with free starting credits, then per-minute billing. Bland’s current pricing references free credits — verify current allowance directly.

What to actually test in your free trial

Voice quality on your specific business name, services, and jargon
Booking accuracy against your real calendar
Caller experience with strong accents or background noise
Handoff behavior when the AI doesn't know an answer
After-hours email and SMS notifications to your team
CRM and calendar integration accuracy
Spam call filtering

Don’t burn your trial on demo conversations. Wire it to a real phone line on a low-risk workflow (after-hours overflow, FAQ-only line) and measure the outcomes.


How we verified this

Every dollar amount on this page was verified against the vendor’s own pricing page or documentation on . Where pricing wasn’t publicly disclosed or where references conflict (RingCentral standalone vs RingEX-bundled, for example), we said so. We don’t invent prices, paraphrase vendor marketing as verified fact, or claim hands-on testing where it didn’t happen.

What we verified

Published plan tiers and pricing

Billing units (per call, per minute, per unique customer, credit-based)

Stated included usage (calls, minutes, customers, credits)

Published overage rates

Publicly stated setup fees

Publicly disclosed transfer fees

Vendor-published integration claims

What we did not verify on this page

Real call accuracy on production calls

Booking accuracy

Latency in live conditions

Hallucination rates

Escalation success rates

Real support responsiveness

Actual BAA negotiation process and timeline

Real-world caller satisfaction

Evidence levels used on this page

Documentation reviewPricing verified against the vendor’s pricing page and documentation. No hands-on testing.
Documentation review + modeled estimateAll-in cost isn’t published as a single figure (Vapi). Calculated from public provider rates by third-party operators.
Needs primary verificationPricing not publicly available or conflicting (Dialzara, RingCentral). Verify directly before use.
Hands-on trial (other pages)Real calls run through the vendor on a defined test protocol. See our medspa review.

Editor of record: Jordan M. Reyes, The AI Agent Report. Last reviewed: . Next scheduled review: August 20, 2026 (quarterly).


AI receptionist cost FAQ

How much does an AI receptionist cost per month?
Most U.S. small businesses pay $79–$249 per month for an AI receptionist that books appointments and integrates with their stack. The cheapest entry plans start at $25–$29/month with hard call caps; premium and hybrid plans run $250–$800/month. Developer voice-agent platforms are priced per minute at $0.07–$0.40 all-in.
How much does an AI receptionist cost per minute?
Managed AI receptionists typically charge $0.25–$0.48 per minute on overage. Developer voice-agent platforms advertise $0.05–$0.09 per minute but real all-in cost (after speech-to-text, LLM, voice generation, and telephony) lands at $0.13–$0.40 per minute. Per-minute sticker rates are almost always lower than the real bill.
What’s the cheapest AI receptionist?
Entry tiers start at $25/month (Sona by Quo, credit-based) and $79/month (Frontdesk billed annually, Goodcall Starter). These plans cap calls or minutes — the cheapest is rarely the cheapest if you exceed the cap. For most small businesses, the cheapest usable plan is closer to $99–$129/month.
Is an AI receptionist cheaper than a human receptionist?
Usually, yes — for the call types AI handles well. A full-time human receptionist costs $3,875–$5,000/month loaded (BLS lists the 2024 median wage at $37,230/year plus benefits and overhead). An AI receptionist that handles routine call coverage runs $79–$249/month. The exception: high-stakes calls requiring empathy and judgment, where human or hybrid services are worth the premium.
Are there hidden fees with AI receptionists?
Yes. Common hidden costs include setup fees ($0–$500+), transfer-time fees ($0.04–$0.05/min on some platforms), failed-call minimums ($0.015 per attempt on Bland), SMS line items ($0.01–$0.05/message), integration tier-ups, BAA add-ons, and annual-only pricing lock-ins. The 10-question quote audit on this page lists what to ask before signing.
Do AI receptionists charge by call or by minute?
Both — and per unique customer, per credit, or per agent are also common. Per-call billing wins when calls average over 2 minutes and volume is consistent (Smith.ai). Per-minute wins when calls are short or volume varies (Retell, Bland, Ruby). Per-unique-customer wins when you have repeat callers with long calls (Goodcall). Flat monthly is best for predictable budgeting (Frontdesk).
Can an AI receptionist book appointments?
Yes, most can — through native integrations with Google Calendar, Calendly, Cal.com, or specific practice-management systems. Booking accuracy depends on the vendor’s calendar logic and how well your workflow is configured.
Does an AI receptionist need human backup?
It depends on your call types. Routine calls (hours, services, basic FAQs, simple booking) don’t need human backup. High-stakes intake (legal, medical triage, crisis), complex insurance questions, and angry-customer escalation benefit from human backup. Smith.ai’s AI Receptionist offers on-demand human handoff; Smith.ai Virtual Receptionists is human-first.
Is an AI receptionist legal?
AI receptionist software can be used legally in the U.S., but outbound calling, SMS follow-up, call recording, AI disclosure, HIPAA/PHI handling, and state-specific rules depend on workflow and jurisdiction. Verify your specific use case with qualified counsel.
Do I need to disclose that callers are speaking with AI?
Legal minimums vary by state and call type. Best practice regardless of legal minimum: have your AI receptionist identify itself as an AI assistant near the start of the call. It pre-empts most disclosure concerns, builds trust, and sets caller expectations correctly. Verify your specific state and use case with qualified counsel.
Can an AI receptionist be HIPAA compliant?
Several vendors publish HIPAA-related documentation and offer BAA execution. A vendor ‘being HIPAA compliant’ is not the same as having executed a BAA with you. Don’t process protected health information until the BAA is signed in writing, and verify scope directly with the vendor.
What should a small business budget for an AI receptionist?
Budget $100–$300/month for a real pilot at typical SMB call volumes (30–150 calls/month). Add 20–30% buffer for overages and add-ons in your first 90 days while you tune the workflow. If your business is regulated (healthcare, legal, financial), add compliance review and potentially a higher plan tier for BAA or SOC 2 documentation.

Find the right AI receptionist for your call volume

You’ve read the cost breakdown, the billing models, the hidden fees, the vendor matrix, and the compliance overlay. The next step is matching your specific call volume, vertical, and stack to the 2–3 vendors that fit. The matching framework takes about 60 seconds. Affiliate relationships are disclosed on every recommendation.


Related reading

This page publishes software buying research, not legal, medical, financial, or compliance advice. Operators should verify regulatory obligations (TCPA, HIPAA, state AI disclosure laws) with qualified counsel before deploying AI agents in regulated workflows.

Page version: 1.0 · Last reviewed: · Next scheduled review: August 20, 2026 · Editor: Jordan M. Reyes

Find My AI Agent Path

60 seconds · No email needed