How to Price an AI Receptionist for Your Agency: Models, Margins and Real Numbers

Price an AI receptionist as a flat monthly package with included minutes and a per-minute overage, not as bare minutes. In the example below, a $299 package for a client who uses 300 minutes leaves 90.0% after platform cost and 63.2% after your support time. Every figure in this post shows its arithmetic, so you can swap in your own numbers.

We build PulsyAI, so the cost side uses our published prices. The same method works on any platform. Run it on ours and on the others before you sign anything.

Key Takeaways

• Your platform cost is a fixed fee plus a per-minute rate. On PulsyAI that is $99 per month with 1,000 minutes included, as published in October 2026.
• Our own homepage margin examples ($200, $400, $700) subtract only the $99 fee for one client. They show 66.9%, 80.2% and 87.6%, not a blanket "80%+".
• A flat package with included minutes and an overage rate is the easiest model to sell and the easiest to protect.
• Support time is the biggest hidden cost. At 2 hours per client per month and $40 per hour (an assumption), a 90.0% margin becomes 63.2%.
• Heavy users are the main risk of flat pricing. A $299 client breaks even at 2,990 minutes per month if every minute costs you $0.10.

What does an AI receptionist cost you to run?

Start with your cost base, because every pricing model is just a different way of recovering it. Here is what PulsyAI publishes as of October 2026, from the homepage and the features page.

ItemPublished figureWhite-Label plan$99 per month with 1,000 minutes included (shown as a limited-time offer)Starter planFlat $0.10 per minute, up to 5 concurrent calls, no custom brandingSub-accountsUnlimited, with no per-seat or per-agent feeModel providersIncluded in platform cost, no bring-your-own-key requirementContractNo contracts, cancel anytime

One gap matters for your math. Our pages don't state the per-minute rate after the 1,000 included minutes. The features-page FAQ describes the price as "$99/month plus $0.10/minute". This post therefore models overage at $0.10 per minute, the Starter rate. Check the rate when you sign up and rerun the tables if it differs.

Two more costs sit outside the platform fee in most setups: carrier minutes and your own time. If you bring your own SIP trunk, Twilio lists US local inbound calls at $0.0085 per minute plus $1.15 per month per number (Twilio US voice pricing, October 2026). Our pages don't say whether carrier minutes are included in the $0.10, so confirm that before you quote a client.

How do the four main pricing models compare?

Agencies usually pick one of four structures. Here is each one with the margin math. All dollar figures are examples, and the $0.10 per minute cost is the assumption stated above.

Model 1: Per-minute resale

You charge the client a rate per minute and pay the platform $0.10. The margin per minute is the price minus the cost.

• At $0.25 per minute: $0.25 − $0.10 = $0.15 margin, which is $0.15 ÷ $0.25 = 60.0%.
• At $0.30 per minute: $0.20 margin, which is 66.7%.
• At $0.35 per minute: $0.25 margin, which is 71.4%.

The $0.25 example matches the one on our homepage, which works out to 60.0%, not 80%. The weakness of this model is the bill. A client who pays per minute sees a variable invoice, and it drops to zero in a quiet month. You also have no recurring base to cover your fixed costs.

Model 2: Flat package with included minutes

You charge a monthly fee that includes a minute allowance, then bill overage above it. The client gets a predictable invoice, and you get recurring revenue. Our homepage shows $299, $499 and $799 as example price points, so this post uses them.

The risk is the heavy user. If a $299 client talks for 1,500 minutes, your marginal cost is 1,500 × $0.10 = $150. That leaves $149, which is 49.8%. The break-even is $299 ÷ $0.10 = 2,990 minutes. An overage rate protects you before you get there. If you charge $0.30 per minute over the allowance against a $0.10 cost, each extra minute earns $0.20, or 66.7%.

Model 3: Setup fee plus retainer

You charge a one-time setup fee for building and tuning the agent, then a monthly retainer. The setup fee pays for your hours, so it shouldn't be treated as margin on the platform.

Example only: a $500 setup fee against 6 hours at $40 per hour ($240) leaves $260, or 52.0%. The assumption is that the build takes 6 hours. Track your real hours on the first three clients and replace the 6.

Model 4: Tiered packages

You offer three packages that differ by minutes and by how much work you do. This is the same flat-package idea with a clearer upgrade path. The table below uses a deliberately conservative cost: it charges every included minute at $0.10, even though the first 1,000 minutes across all your clients are covered by the $99 plan.

PackagePriceIncluded minutesPlatform cost (minutes × $0.10)Margin after platformSupport hours × $40Margin after supportAnswering$299300$30$269 (90.0%)2 h = $80$189 (63.2%)Answering plus booking$499600$60$439 (88.0%)3 h = $120$319 (63.9%)Multi-location$7991,500$150$649 (81.2%)5 h = $200$449 (56.2%)

The support hours and the $40 hourly rate are assumptions for illustration. They are not benchmarks. The point of the table is the last column: the platform cost is small, and your time decides the margin.

Why do the margins on our homepage look higher than yours?

Our homepage shows $200 margin at $299, $400 at $499 and $700 at $799. Those figures are the price minus $99, which is one client absorbing the whole platform fee.

PriceHomepage marginArithmeticPercent$299$200$200 ÷ $29966.9%$499$400$400 ÷ $49980.2%$799$700$700 ÷ $79987.6%

So the "80%+" line on the page only holds for the two higher price points, and only if one client's usage fits inside the 1,000 included minutes. Use these as a ceiling for a single client, not a forecast.

What happens when you add clients?

The fixed $99 fee spreads across your clients, and the 1,000 included minutes become a shared pool. Here is a worked example with 10 clients at $299 each, so $2,990 in monthly revenue. Platform cost is $99 plus $0.10 for every minute above 1,000.

Minutes per clientTotal minutesPlatform costMargin after platformPercent of $2,9902002,000$99 + 1,000 × $0.10 = $199$2,79193.3%3003,000$99 + 2,000 × $0.10 = $299$2,69190.0%6006,000$99 + 5,000 × $0.10 = $599$2,39180.0%1,00010,000$99 + 9,000 × $0.10 = $999$1,99166.6%

Now add your time. Take the 300-minute row and assume 2 hours of support per client at $40 per hour. That is 10 × $80 = $800. The margin becomes $2,691 − $800 = $1,891, and $1,891 ÷ $2,990 = 63.2%. This is the number to plan around.

What will your clients compare you to?

Your client doesn't know your costs. They know what other answering options charge. Two public price points, both read from the vendors' own pages in October 2026:

• Goodcall lists Starter at $79 per month for 100 unique customers, Growth at $129 for 500, and Scale at $299 for 2,000 (goodcall.com/pricing).
• Smith.ai lists Starter at $300 per month for 30 calls, with $11.50 per call over 30 (smith.ai/pricing). The page doesn't say how the prices differ between its AI-first and human-first options. $300 ÷ 30 calls is $10 per included call.

A self-serve product at $79 puts a ceiling on how much a client will pay for software alone. Your $299 package has to sell the managed part: setup, tuning, reporting and someone to call. If the client only wants software, you will lose on price. If they want it handled, price on the hours you save them.

Where does margin quietly disappear?

1. Heavy users. One client at 1,500 minutes costs you 5 times a client at 300. Set an overage rate before you send the first invoice.
2. Support time. Every prompt change, number port and "why did it say that" call is unpaid unless the package says otherwise. Write down what is included.
3. Telephony. If carrier minutes aren't in your platform rate, add them. At $0.0085 per minute on Twilio, 3,000 minutes cost 3,000 × $0.0085 = $25.50, small but not zero.
4. Churn. Illustrative only: at 5% monthly churn, a client stays 1 ÷ 0.05 = 20 months on average. At $189 margin after support, that is about $3,780 per client. Use your own churn once you have three months of data.

How do you set your own price? A five-step worksheet

1. Write down your platform cost per minute and your fixed fee. Confirm the overage rate with the vendor.
2. Estimate minutes per client. Use the first month of real data as soon as you have it, and keep the 300-minute guess as a placeholder only.
3. Estimate your support hours per client and your hourly cost. Be honest. A low number here is the most common pricing mistake.
4. Pick a package price, then compute margin after platform and after support using the formulas above.
5. Check the price against what the client can buy elsewhere, then set an overage rate at least 2 times your minute cost.

The formula is: margin = price − (minutes × cost per minute) − (support hours × hourly cost). Every table in this post is that formula with different inputs.

Where does PulsyAI fit, and where doesn't it?

The pricing math works best when you have several clients, because the fixed fee spreads out and Stripe rebilling handles the invoices. Our features page describes Stripe Connect rebilling with a base fee, included minutes and overage rates, with payments going to your own Stripe account. If you plan to resell through other resellers, see multi-tier white-label. If you are weighing the partner program first, start with partners.

It fits less well if you have one or two clients with very high call volume. In that case the per-minute cost dominates, and you should compare the minute rate on every platform you are considering.

Disclosure: PulsyAI publishes this post and sells a product that competes with the vendors named above. Competitor prices were read from their own pages in October 2026 and can change. Margin figures are arithmetic on the stated assumptions, not results from any customer.