If you've been considering an AI voice agent—for handling incoming calls, qualifying leads, or scheduling appointments—you've probably wondered what it actually costs. Not the headline number that gets advertised, but what you'll pay each month.
Retell is one of the leading voice AI platforms, and understanding its true cost is essential for deciding whether voice AI makes financial sense for your oil & gas company, manufacturing facility, or property management operation.
The real cost: $0.13 to $0.31 per minute for a fully working setup. The advertised $0.07/minute covers only the core platform infrastructure—you'll need voice, LLMs, and telephony on top, which push the true cost into that range. Users report real costs 30–60% above the advertised minimum, largely due to the developer work needed to build and maintain the integration. There are also free credits to start.
Breaking down the costs
The base rate
Retell uses usage-based pricing at $0.07 per minute for the core API—the rate you'll see advertised. But a fully functioning voice agent involves several cost layers:
Speech recognition (ASR) — converts spoken audio to text
Large language model (LLM) — interprets intent and generates responses
Text-to-speech (TTS) — converts responses back to spoken audio
Telephony — connection to the public phone network
Platform fees — managing call state and orchestration
The $0.07 base covers platform infrastructure only. With dependencies included, the full cost typically runs $0.13 to $0.31 per minute.
What drives the final number
For a typical setup with voice infrastructure, TTS, LLMs, and telephony, the exact cost depends on:
Which LLM you use — more capable models cost more
Which TTS voice you choose — premium voices may cost extra
Call volume — higher volumes may qualify for discounts
Add-ons — analytics, multilingual support, and compliance features
One point in your favor: Retell doesn't charge for calls that fail to connect—you only pay for connected conversation time.
Free credits to start
Retell offers $10 in free credits for new users—enough to test the platform, build an agent, and evaluate performance before committing to paid usage.
Where voice AI fits, by industry
The same per-minute economics apply across industries—what changes is call volume, call length, and how much premium handling each use case needs. Here's where voice AI tends to earn its keep:
HVAC
After-hours emergency calls, maintenance scheduling, and routing urgent no-heat/no-cool requests. The biggest win is coverage: emergency calls get answered at 2 a.m. without the owner picking up the phone, and routine scheduling stops eating daytime staff hours.
Construction
Inbound lead qualification, project-scope intake, and subcontractor or change-order calls. Every lead gets an immediate response instead of going to voicemail, and unqualified inquiries are screened before they reach the project manager. Longer, more detailed calls here tend toward the higher end of the rate range.
Law / legal
Client intake, case-detail gathering, conflict screening, and consultation scheduling. Legal calls often need premium voice and careful handling—so they sit at the higher end of the per-minute range—but they capture after-hours leads, when many legal inquiries actually come in, and route qualified matters to an attorney.
Dental
Appointment booking, rescheduling, insurance pre-questions, and recall reminders. These are short, high-volume, simple calls—the lower end of the rate range—that free the front desk from routine scheduling during patient check-ins and ensure no booking request goes unanswered at lunch or after close.
Industrial services
Service requests, equipment-issue logging, field-technician coordination, and callbacks. Requests are captured and triaged around the clock, and coordination happens without tying up office staff.
Transportation
Driver check-ins, load-status inquiries, scheduling, and customer ETA calls. High call volume is handled without scaling dispatch headcount, and routine status questions are answered instantly instead of queuing.
The pattern across all of these: voice AI pays off most where call volume is high, coverage needs extend beyond business hours, and a meaningful share of calls are routine enough to handle without a transfer. The right setup—and the right per-minute tier within that $0.13–$0.31 range—depends on your specific mix of volume, complexity, and how much human handoff you need.
The business case for voice AI
Priced correctly, voice AI saves significantly against human staffing—especially in:
High-volume environments, where every call costs staff time
After-hours coverage, where you need availability but not a full-time hire
Initial qualification, where you're screening before personal attention
The key is comparing the all-in cost to your current alternatives—not just the platform rate, but the value of the outcomes you're after.
Key considerations before you commit
Volume — higher call volumes improve ROI; the economics at 100 calls/month differ sharply from 1,000.
Complexity — simple workflows (scheduling) automate easily; complex diagnosis is harder.
Handoff — what share of calls need a human? Voice AI works best when it can qualify and route without transferring.
Integration — seamless connection to your CRM and scheduling tools is essential.
Measurement — track missed-call reduction, response time, and qualification accuracy to know if it's delivering.
The investment isn't just the per-minute cost. It's the staff time saved and the customer interactions you capture that you'd otherwise miss.
