top of page
logo

Your Next Customer-Service Call May Be Answered by an AI That Costs 3.5 Cents a Minute

sheen6869
2 hours ago
5 min read

By Shakeel Ahmed — AI Growth Lab | November 13, 2026



Imagine calling a small business and getting a helpful answer in under a second, at any hour, for 3.5 cents a minute. That is the promise Retell made on November 12, 2026, when it launched Tell-1, a language model built specifically for live contact-center calls, along with a new Dialer product for outbound campaigns. The model is priced at 0.035 dollars per minute in beta. For a small business owner who has ever missed a call, lost a customer to voicemail, or paid for a call center they could not afford, this is not a tech story. It is a hiring decision.


WHAT TELL-1 ACTUALLY IS


Retell is an AI voice agent platform for contact centers. Tell-1 is its new language model, built for the complexity of live phone calls. A phone call is harder than a chat conversation. People interrupt each other. They mumble. They change their minds mid-sentence. They get angry. A model built for text has to be adapted to survive all of that in real time, and that is what Tell-1 is designed to do.


Two numbers matter most. First, latency. On Retell's live benchmark, Tell-1 is currently responding in under 100 milliseconds, and it is built to stay under 300 milliseconds even at peak call volumes. A response under 100 milliseconds feels instant. A response over a second feels like talking to a machine that is thinking too hard. Speed is not a nice extra in voice AI. It is the whole product.


Second, cost. An extra 5 to 10 cents per call adds up quickly across millions of calls a month, and Retell built Tell-1 to deliver the lowest possible latency and cost without sacrificing call quality. On Retell's own vCX Voice AI Benchmark, Tell-1 posted the highest overall score of any model tested. That is the company's own benchmark, so treat it as a claim rather than an independent verdict, but the direction is clear: the company that sells voice agents just made its own voice agent cheaper and faster at the same time.


THE FULL PRICE, HONESTLY STATED


The 3.5 cents a minute is the price of the Tell-1 model in beta. It is not the whole call. Retell's pricing page breaks the call into parts: voice infrastructure at 0.055 dollars a minute, the language model, text-to-speech from about 0.015 dollars a minute, and telephony at 0.015 dollars a minute on standard rates. Put together, a full AI voice call on Retell typically works out at about 0.07 to 0.31 dollars a minute depending on the model and add-ons, plus the phone line. A phone number costs 2 dollars a month. The first 20 simultaneous calls are free.


That is still a small number compared with a human agent. A human support worker costs several dollars per call once you count wages, training, and management, and a human can only take one call at a time. An AI agent can take a hundred calls at once, at 3 in the morning, without a shift change. A five-minute customer call on Tell-1 costs about 18 cents for the model. The same five minutes of a human agent's time costs many times more, and the human was probably asleep.


But the honest comparison is not AI versus human on price. It is AI versus a missed call. A missed call costs a small business the whole customer. An AI that answers, books the appointment, and takes the message is not replacing a human. It is replacing voicemail.


WHAT THE DIALER ADDS


Retell also launched Retell Dialer on the same day, the first dialer built for voice agents, to run outbound campaigns across AI voice and SMS agents. Where Tell-1 handles decisions during a live conversation, Dialer turns the outcome of that conversation into the next action.


Here is how that works in practice. A requested callback replaces the next planned attempt and resumes the same conversation at the time the contact asked for. A qualified prospect moves to a specialist agent. Someone interested but not ready receives an SMS follow-up from a nurture agent. Dialer also handles contact screening, local presence and calling windows, number rotation and spam monitoring, configurable retries, and CRM integration.


In plain words: the AI does not just answer the phone. It runs the whole follow-up. For a small business doing appointment reminders, payment follow-ups, or lead outreach, that is the difference between a tool and a system.


WHO SHOULD TRY IT FIRST


Three kinds of small business should look at this now.


First, businesses that lose money to missed calls: clinics, salons, repair services, restaurants taking bookings. If your phone rings while you are with a customer, an AI that answers and books the appointment pays for itself in a day.


Second, businesses with predictable, repetitive calls: appointment confirmations, delivery updates, opening hours, order status. These are the calls an AI handles best, because the answers are known in advance.


Third, businesses doing outbound follow-up by hand. If someone on your team spends hours a day calling back leads or reminding customers, Dialer turns that into a system that runs itself, with a human stepping in only when the prospect is qualified.


Who should wait? Anyone whose customers call with complex, emotional problems that need a human ear. A billing dispute, a complaint, a crisis: these are the calls where a fast AI answer can feel like an insult. The smart setup is a hybrid. The AI answers first, handles the routine, and hands the hard calls to a human immediately, with a summary of what the caller already said. Nobody likes repeating their problem twice.


THE TRUST QUESTION


There is one question every business owner should ask before putting an AI on the phone: will my customers accept it? The answer depends on how honest you are. An AI that pretends to be human is a trust problem waiting to happen. An AI that says, clearly, "I am the automated assistant, I can help you book or answer questions, and I can connect you with a person any time," is a service.


Voice AI is now good enough that the question is no longer whether it works. The question is whether you use it to respect your customers' time or to avoid them. Customers can tell the difference, and they remember it.


A year ago, a voice agent that could hold a real conversation cost real money and needed real engineering. Today the model costs 3.5 cents a minute and the platform is pay-as-you-go, with no minimums and no contracts. The same week, Google launched a coding agent model at 1.50 dollars per million tokens and a decision model at 10 cents. The pattern is the same everywhere: the cost of hiring an AI worker keeps falling, and the businesses that move first are the ones whose AI handles the routine while humans handle the relationships.


Your next customer-service call may be answered by an AI that costs 3.5 cents a minute. The real question is whether your business will be the one answering, or the one losing the call to someone who is.


Sources


Retell, "Retell Launches Tell-1, a Language Model Built for the Complexity of Live Contact Center Calls," Business Wire, November 12, 2026: https://www.businesswire.com/news/home/20261112258243/en/

Retell AI pricing page: https://www.retellai.com/pricing

AY Automate, "Best AI Receptionist 2026: 7 Compared on Price," updated November 6, 2026: https://www.ayautomate.com/blog/best-ai-receptionist


ChatGPT just became a storefront — your business can now sell inside the chatbot: https://dev.to/shakeelahmed/chatgpt-just-became-a-storefront-your-business-can-now-sell-inside-the-chatbot-4nip

Meta's new AI costs 10 cents — and that's why your next app will have an agent in it: https://dev.to/shakeelahmed/metas-new-ai-costs-10-cents-and-thats-why-your-next-app-will-have-an-agent-in-it-ej5

Comments

bottom of page