AI Receptionist for US Small Business: Voice, Chat, SMS Cost Line

What an AI receptionist actually does in 2026 (spoiler: more than answering)

A modern AI receptionist for a US small business handles four distinct jobs across three channels:

Jobs: (1) answer the call/chat/text within seconds; (2) qualify the caller — new client, existing client, sales inquiry, support issue, or wrong number; (3) execute the appropriate action — book an appointment in the calendar, send an intake form, transfer to a human agent, or take a message; (4) log the interaction into whatever CRM the business already uses.

Channels: voice (traditional phone line), text/SMS (Twilio-backed short code or long code), and chat (WhatsApp Business, web widget, or Facebook Messenger). A good AI receptionist supports at least voice + SMS + one chat channel.

What's changed in 2026 vs the old answering service model:

• Voice is genuinely conversational, not menu-driven. GPT-4-class LLMs behind voice tools (like Retell, Vapi, Bland, Air.ai) can hold a natural back-and-forth without scripted trees. The caller doesn't hear 'press 1 for sales, press 2 for support' — they say 'I want to book a cleaning for Tuesday' and the AI responds like a human.
• Booking is direct. The AI reads the calendar (Google Calendar, Calendly, Acuity, Square Appointments, or the PMS for medical/dental practices) and creates the appointment in real time, not by taking a message for a human to book later.
• Cross-channel handoff works. A caller who leaves a voicemail can get a followup SMS with a link to complete booking online. A WhatsApp inquiry can transfer to a phone call when the customer explicitly asks.

The old answering service model still exists (Ruby Receptionists, PATLive) and still fits businesses where the human answering the call IS the product experience (bespoke high-touch service). For most small businesses handling common inquiries, the AI receptionist model does the same job for meaningfully less per month while handling more channels.

The four vendor categories and when each fits

The US AI receptionist market in 2026 splits into four categories:

1. Voice-first AI answering services. Retell AI, Vapi, Bland AI, Air.ai, Synthflow. Purpose-built voice AI that answers phone calls. Pricing typically per-minute of call time — approximately $0.05-$0.30 per minute depending on vendor and voice model quality. A dental practice getting 200 calls/month averaging 3 minutes each pays roughly $30-180/month in AI cost, plus a phone number rental ($1-2/month for Twilio) and any BSP fees. Best for businesses whose primary customer channel is inbound phone.

2. Chat-first with voice bolt-on. Intercom Fin, Ada, Drift with voice enabled, Zendesk with voice. These platforms started as chat and added voice. Pricing typically per-user or per-conversation ($100-500+/month for the platform). Best for businesses that already need a full CRM with support ticketing and want an AI layer.

3. Vertical-specific AI receptionists. Solutions purpose-built for one vertical — dental (Weave, Solutionreach, Dentrix Ascend), medical practices (Luma Health, Klara, RingRx), law firms (Clio Grow, MyCase), home services (Housecall Pro, Jobber). Priced as part of the vertical suite ($100-400/user/month typically). Best for businesses already using the vertical's operating system.

4. WhatsApp Business platform BSPs with AI response. WATI, respond.io, 360dialog, Sleekflow. Chat-only (no voice), typically $29-99+/month. Best for businesses whose customers actually contact them via WhatsApp — meaningful in US markets serving immigrant SMBs (US-Latino, US-Asian consumer segments).

Which fits a specific business: If most customer contact is inbound phone, a voice-first vendor (Retell, Vapi) plus SMS on the same number is the cleanest stack. If customer contact is mixed across web chat + phone + email, a chat-first platform (Intercom Fin) covers the whole stack. If the business runs on a vertical suite (a dental practice on Weave), the AI receptionist in the suite is usually the right pick because it integrates with the calendar, patient records, and billing directly.

Real cost breakdown: what a US small business actually pays

Concrete modeled scenarios for a US small business considering an AI receptionist:

Solo service business (single owner, 100 calls/month, mostly bookings). Voice-first vendor at $0.10/minute × average 3-minute call = $30/month AI cost. Twilio phone number $1-2/month. WhatsApp Business App (free) for text follow-up. Total: $30-40/month. Compare to Ruby Receptionists at $250-350/month for equivalent human-answered coverage.

Dental or medical practice (small — 500 calls/month, appointment booking + patient questions). Vertical vendor (Weave, Luma Health) at $150-300/month per operatory or per user, covers voice + SMS + patient recall + review requests. AI response is bundled. Standalone voice AI (Retell) + separate patient management would cost $150-200/month combined but requires integration work. Total: $200-400/month bundled or $150-250 pieced together.

Home services company (plumber, electrician, HVAC — 300 calls/month with dispatch decisions). Housecall Pro or Jobber ($75-200/month) with built-in booking + dispatcher AI, or Retell + a Zapier integration to the field service management tool ($50-100/month combined). Total: $75-300/month depending on integration depth.

Law firm (small — 50 inbound calls/month with strict privilege considerations). Clio Grow ($59-129/user/month) with AI intake or Retell configured with attorney-review-required flags before booking consultations ($30-60/month). Law firms often prefer human intake for privilege reasons — hybrid AI-triage + human-handoff is common. Total: $60-200/month.

E-commerce or D2C brand (moderate — 200 support conversations/month across email + chat + WhatsApp). Intercom Fin AI ($99-299/month) or Zendesk with AI ($55-115/agent/month) if already on those platforms. WATI ($29/month) + Twilio SMS if WhatsApp-first. Total: $30-300/month depending on channel mix.

Enterprise mid-market (50+ agents, high call volume). Custom pricing negotiated with Five9, Genesys, RingCentral MVP with Nice or Talkdesk. Total: $2000-15000+/month — not the SMB segment this guide addresses.

The consistent lesson: the AI receptionist market has broken the old $200-500/month answering-service price floor. Solo service businesses can get functional AI answering at $30-50/month. What determines total cost is not the AI itself — it's the integration depth with the business's existing calendar, CRM, and vertical software.

TCPA, HIPAA, and state privacy: the US compliance floor

US small businesses using AI receptionists face three main federal or state regulatory frameworks:

TCPA (Telephone Consumer Protection Act) and FCC rules. TCPA governs autodialed calls, text messages, and prerecorded messages to consumer phones. In 2023-2024, the FCC clarified that AI-generated voice on outbound calls counts as an 'artificial or prerecorded voice' requiring prior express written consent from the recipient. Inbound calls where the caller reaches the business are exempt from TCPA consent rules (the caller initiated contact). Outbound SMS from an AI-driven follow-up sequence needs opt-in per TCPA. Practical implication: inbound AI receptionist is TCPA-safe by default; outbound AI SMS or voice campaigns need documented consent.

HIPAA (Health Insurance Portability and Accountability Act). Applies to covered entities — health plans, healthcare clearinghouses, and most healthcare providers who transmit health information electronically. Business Associates (vendors handling PHI on behalf of covered entities) must sign a Business Associate Agreement (BAA). For AI receptionists used in dental, medical, mental health, or veterinary practices, the vendor must offer a BAA and structure the service to keep PHI encrypted at rest and in transit. Vendors with published BAA support: Weave (dental), Luma Health (medical), Klara (medical). Voice-first AI vendors (Retell, Vapi) vary on BAA availability — verify before contracting if the practice handles PHI.

State privacy laws. The state-privacy landscape expanded fast in 2023-2025:

• California — CCPA (California Consumer Privacy Act) and CPRA (California Privacy Rights Act) apply to businesses meeting revenue or consumer-count thresholds. Consumer right to know, delete, correct, and opt out of sale/sharing.
• Virginia — CDPA (Consumer Data Protection Act).
• Colorado — CPA (Colorado Privacy Act).
• Utah — UCPA (Utah Consumer Privacy Act).
• Connecticut — CTDPA (Connecticut Data Privacy Act).
• Additional states adopted or adopting: Iowa, Indiana, Tennessee, Montana, Texas, Oregon, Delaware, New Hampshire, New Jersey, Kentucky, Rhode Island (verify current status per state AG).

Most state laws apply to businesses above enterprise thresholds (e.g., $25M revenue in California, or processing personal data of 100,000+ consumers). Micro and small businesses often fall below the enterprise thresholds but still owe reasonable security for personally identifiable information under state consumer protection laws.

Practical AI receptionist compliance checklist:

1. Written business associate agreement with the AI vendor if healthcare-adjacent.
2. Documented consent capture for outbound AI voice or SMS campaigns.
3. Privacy policy on the business website disclosing AI receptionist use and vendor sub-processors.
4. Reasonable security for call recordings and transcripts (encrypted at rest, access-controlled).
5. Retention policy — do not keep call recordings indefinitely without a documented purpose.

The compliance floor is not high for a solo service business or e-commerce brand. It gets meaningful for healthcare, legal, and financial services — those verticals should consult specialty counsel before contracting.

When to hire a human receptionist instead

AI receptionists are not the right answer for every business. Human receptionists still win in specific situations:

High-touch premium service. A luxury wedding planner charging $50,000 per event, a bespoke estate lawyer, a concierge medical practice — the first phone contact IS the product experience. AI voice sounds good in 2026 but a caller expecting a warm human welcome will detect and downgrade the vendor. Book a human answering service.

Very low call volume (fewer than 30 calls/month). The setup effort for AI receptionist integration exceeds the payoff at extremely low volume. A part-time human (spouse, business partner, virtual assistant service like Time Etc or Belay) is often simpler and roughly equivalent in cost at this volume.

Highly regulated intake where privilege or fiduciary duty attaches. Law firm intake carries attorney-client privilege implications; investment advisor intake carries fiduciary duty. The safer default is human intake with attorney or advisor review before commitment. AI triage can still front the call for basic routing, but the substantive intake is human.

Complex diagnostic or emotional conversations. Mental health crisis line, hospice inquiry, complex insurance dispute. AI voice in 2026 handles procedural conversations well; it does not yet handle emotional complexity, ambiguity, or crisis with human warmth. Route these directly to a human.

Very high average call value. A B2B enterprise sales team where each inbound call could close a $100,000 deal — the return on human handling is orders of magnitude higher than the cost. Skip AI, staff human SDRs.

When AI wins clearly: appointment booking for a hair salon, dental practice, dog groomer, physical therapist, tutoring service, home cleaning; order status questions for e-commerce; lead qualification for home services (plumber, HVAC, electrician); FAQ handling for gyms, yoga studios, coaching businesses; SMS reminders and confirmations across every vertical.

Implementation reality: what actually takes time

The cost breakdown above is the running cost. The setup effort determines whether the AI receptionist actually works in practice. What takes time:

Knowledge base curation (2-8 hours). The AI needs to know the answers to common questions — hours of operation, service prices, insurance accepted, parking situation, cancellation policy, service area. The business owner or manager typically writes this. Copying an existing FAQ page saves time but often reveals gaps.

Calendar and CRM integration (1-4 hours). Connecting the AI to the actual booking system takes time — OAuth setup, mapping service types to appointment slots, defining buffer times, blocked hours, and holidays. Some vendors (Weave, Housecall Pro) have this built in for their vertical; standalone AI vendors need manual connection.

Voice tuning (1-3 hours). Testing the AI voice against actual call scenarios — the caller with a strong accent, the caller who talks over the AI, the caller who says something ambiguous. Adjusting the prompt and the fallback rules until it handles the top 20 patterns well.

Compliance setup (variable — 2 hours for e-commerce, 10-30 hours for healthcare). BAA signing, privacy policy update, opt-in flow for outbound, retention policy, consent capture. Healthcare and legal verticals take meaningfully longer.

Team training (1-2 hours). Existing staff needs to know when the AI hands off and how to pick up the conversation. Handoff scripts, dashboard access, escalation paths.

Total realistic setup: 8-20 hours for a typical small business. Businesses that skip this setup and expect the AI to work out-of-the-box report mediocre results and often churn within 60 days. Businesses that invest the setup time typically report 70-85% call containment (AI resolves without human handoff) within the first month.

BossBot is one option to evaluate — WhatsApp-first with SMS integration via Twilio, calendar integration with Google Calendar and Calendly, US-market-aware compliance posture (TCPA-conscious outbound flow, BAA available on request for healthcare-adjacent use), multi-agent inbox with handoff, and English NLU. Public pricing at bossbot.uk/pricing. Free 7-day trial without payment method.