What actually moves the price of an AI voice agent.
AI voice agent pricing has two parts: a build cost — telephony, qualification logic, handoff design — and a running cost that scales with call volume. The AI model fees are usually the smallest line in both. This guide explains what genuinely moves each part, without invented per-minute rates, and how the number gets fixed for your call volume in writing.
We publish no rates or per-minute figures because we do not bill by the guess. Every agent is priced once the scope and call volume are known: a paid discovery sprint, then a fixed, transparent quote in writing.
Five drivers move every voice-agent price.
Call volume and usage fees
The running cost of a voice agent scales with minutes: model and telephony fees are metered per call, so a hundred calls a week and a thousand calls a day are different economics. Volume also changes the engineering — LeadTrack AI's platform was hardened across 100K+ live calls, and surviving that volume is a design input, not an afterthought.
What moves it- Calls per day and average minutes per call
- Model and telephony fees metered at your volume
- Peaks — the campaign spikes the platform must absorb
Where the number landsUsage fees are estimated for your actual call volume in the discovery sprint — a forecast you see before committing, not a meter you discover after.
Telephony and latency engineering
A voice agent lives or dies on the phone line. Real-time audio, carrier integration and keeping response gaps inside natural conversational rhythm are hard engineering — voice AI tolerates no awkward pause, and prospects hang up on robots. This layer, not the model, is where most of LeadTrack's build budget went.
What moves it- Real-time audio pipeline and the latency budget
- Carrier integration and number provisioning per market
- Interruptions, accents and tangents handled mid-call
Where the number landsScoped against your markets and telephony stack in the discovery sprint — the latency bar goes into the quote as an engineering commitment.
Qualification logic and conversation scope
What must the agent find out, and how many ways can the conversation go? A single well-defined qualifying conversation costs a fraction of an open-ended assistant; every added intent, script variant and language widens the scope again. On LeadTrack, per-tenant scripts and intent scoring are tuned to each customer's pipeline definitions — that logic is the product.
What moves it- Intents and outcomes the agent must handle end to end
- Script variants, languages and voices
- Intent scoring tuned to your pipeline definitions
Where the number landsThe discovery sprint fixes exactly which conversations v1 owns — and the quote prices that line, in writing.
Human-handoff design
The most consequential design decision is when the agent stops talking. Live transfer to a rep with the transcript and intent score attached, warm-handoff rules, escalation paths for conversations off the script — designed well, humans join exactly where the conversation earns it; designed badly, the agent burns the leads it was built to save.
What moves it- When and how calls route to humans, live
- Transcript and context attached to every handoff
- Escalation rules for conversations off the script
Where the number landsHandoff rules are designed with your sales team in the discovery sprint and built into the fixed scope — not bolted on after the first lost lead.
Tuning after launch
A voice agent is never finished — it is tuned. Real calls surface edge intents no script predicted; models update; qualification criteria evolve with your pipeline. LeadTrack's agents kept improving because every call feeds outcome tracking and behavioural evaluation — that loop is a running budget line, and an honest pricing guide says so.
What moves it- Transcript review and evaluation cycles
- Script and model updates as real traffic teaches
- Outcome tracking tied to your pipeline, not vanity metrics
Where the number landsQuoted as its own transparent care line alongside the build, so the tuning loop is planned — never an emergency invoice.
From unknown to a fixed quote, in three steps.
Paid discovery sprint
We scope the agent together: the conversations it owns, your escalation and handoff rules, telephony and CRM integrations, and expected call volume. You pay for the sprint because the output has standalone value — a scope you could take anywhere.
Fixed, transparent quote
The sprint ends in a number, in writing: a fixed build budget with fixed milestones, plus usage fees estimated at your call volume so the running cost is visible from day one. If scope changes later, the quote changes transparently with it, and you approve the difference first.
Ongoing tuning
After launch the tuning loop begins: transcript review, script evolution and model updates, planned as care rather than emergency invoices. Production voice agents earn trust the way LeadTrack AI did — call by call, measured.
No open-ended day rates, no surprise invoices. 100% of the code, IP and infrastructure transfers to you, and we sign an NDA on request.
What a production voice agent actually took.
LeadTrack AI — the platform behind the price
For LeadTrack AI (Australia) we engineered the full voice-agent platform: natural, interruptible conversation, instant auto-dialling, per-tenant qualification logic, live human handoff and call analytics. The price of a voice agent is the price of that platform working — proven across more than a hundred thousand live calls, with conversion up 38%.
The ops layer — where AI running budgets live
The cost lesson generalises across our AI portfolio: on FeelEat's demand-forecasting platform the model took weeks, while the pipelines, drift monitoring and operational views around it took months. Voice agents follow the same curve — the evaluation and tuning layer around the model is where the running budget lives, and where the value holds.
Questions teams ask about voice-agent pricing
Coloque-o num briefing. Um engenheiro sénior — não um vendedor — responde no prazo de um dia útil.
Q.01How much does an AI voice agent cost?
There are two honest numbers, and neither exists before scoping: a build cost driven by telephony, qualification logic and handoff design, and a running cost that scales with call volume. Our answer is a process: a paid discovery sprint that ends in a fixed build quote plus usage fees estimated at your volume, in writing.
Q.02Why don't you publish a per-minute rate?
Because a per-minute rate hides the two things that actually decide your economics: what the platform around the model must do, and what your call volume really is. A rate honest enough to cover both would be too wide to mean anything. We estimate usage at your volume and fix the build in writing instead.
Q.03Isn't the AI model the main cost?
No — model fees are usually the smallest line. The budget lives in the engineering around the model: telephony and latency, qualification logic, human handoff and the evaluation loop that makes the agent safe with real callers. That platform is where LeadTrack AI's budget went, hardened across 100K+ live calls.
Q.04What does call volume change?
Almost everything on the running side: model and telephony fees are metered per call, and volume determines the infrastructure the platform needs and the pace at which the tuning loop learns. The discovery sprint estimates fees at your expected volume, with peaks named, so the meter is never a surprise.
Q.05What happens when a call goes off the script?
That is a design requirement, not an edge case: escalation rules route the conversation to a human, live, with the transcript and intent score attached. On LeadTrack, humans step in exactly where the conversation earns it — handoff design is part of the fixed scope, and one of the drivers that genuinely moves the price.
Q.06What does the agent cost after launch?
Two lines, both visible up front: usage-based model and telephony fees at your call volume, and the tuning loop — transcript review, script evolution, model updates — quoted as an ongoing-care plan. Voice agents that skip the second line stop improving; ours are priced so the loop is planned.
Q.07How do we start?
With a paid discovery sprint: we scope the conversations, handoff rules, integrations and call volume together, and you receive a fixed, transparent quote before any build begins.
Tem um problema que vale a pena
resolver bem?
Conte-nos sobre o seu produto, o seu prazo e as suas limitações. Responderemos no prazo de um dia útil com uma avaliação honesta sobre adequação, âmbito e a equipa certa para o projeto.