Connected-Call Pricing vs Per-Minute Pricing for AI Calling: How to Compare

DialSwarm team · October 17, 2026 · 6 min read

Why pricing models matter more than the headline rate

Shopping for AI calling tools, you will see rates that look easy to compare: a few cents per minute here, a monthly plan there, a price per call somewhere else. They are not easy to compare. The unit changes what you pay for, and it changes how your team behaves once the tool is running.

A team on a model that bills for every attempt starts cutting follow-up. A team on a model that bills by the minute may rush good conversations. A team on a flat plan may underuse it. This article explains the models, shows real published prices, and gives you a way to compare them on your own numbers.

The main models in the market

Per minute. You pay for time on the line. Depending on the vendor, that may include ringing, voicemail, hold, and transfer time, and telephony may be billed separately.

Per minute plus platform fee. A monthly fee gets you a lower per-minute rate or more features.

Flat monthly tiers. A set price per month for a bundle of features or calls.

Per seat. A price per user, common for sales tools.

Per connected call. You pay when a call connects with a person. Unanswered dials are not billed.

Quote only. Pricing after a sales conversation, often with contracts or seat minimums.

What vendors actually publish

Here are examples from vendor pricing pages, checked on Oct 17, 2026. Prices change often, so confirm on the live pages before deciding.

• Bland: $0.14 per minute on its free plan, or $0.12 per minute plus $299 per month on its Build plan. Telephony and transfer minutes are extra. Bland pricing
• Retell AI: $0.07 to $0.31 per minute, no platform fee, telephony extra. Retell pricing
• Vapi: $0.05 per minute for hosting plus provider costs. Vapi pricing
• Synthflow: enterprise only, with contracts from $30,000 per year. Synthflow pricing
• Orum and Nooks: quote only. Orum lists a 9-seat minimum. Orum pricing, Nooks pricing
• DialSwarm: connected calls only, so unanswered dials don't count. You get 10 free calls with no card, and current plans are on the pricing page.

None of these models is wrong. They reflect different products and buyers, which matters more than the rate itself.

What each model rewards

Per-minute pricing ties cost to time. That works well when calls are short and most attempts connect. It also means a long, productive conversation costs more, and depending on how unanswered time is counted, persistence may cost more too. Read the fine print on ringing, voicemail, minimum increments, and transfers.

Flat tiers reward steady, predictable usage. They can be expensive in slow months and limiting in busy ones.

Per-seat pricing rewards adding usage per rep. It can be costly for teams with many occasional users.

Per-connected-call pricing ties cost to conversations. Retries that do not connect add nothing, so following up properly does not raise your bill. A short connected call costs the same as a long one, which is a tradeoff to understand.

Quote-only pricing can bring custom terms but makes side-by-side comparison slower, and minimums may not suit small teams.

A formula you can run on your own numbers

Use letters, then plug in your own data. We are not giving sample connect rates, because they vary too much by industry and lead source to be useful.

• L = leads you call per month.
• A = attempts per lead in your sequence.
• C = share of attempts that connect.
• M = average billed minutes per connected call.
• U = average billed minutes per unanswered attempt (zero if the vendor does not bill them).
• T = telephony cost per minute, if billed separately.

Per-minute cost is roughly: (L × A × C × M + L × A × (1 minus C) × U) × (rate + T), plus any platform fee.

Connected-call cost is roughly: L × A × C × price per connected call. Or your seat plan, if you are on one.

The bigger A is and the smaller C is, the more the two models can differ. That is exactly the situation of a team doing proper follow-up on inbound leads. Get C from your own call history, or from a small test.

How behavior changes under each model

Pricing shapes habits, so think a few months ahead. Under per-minute billing, managers start watching average call length, and reps may be nudged to wrap up conversations sooner. Under per-attempt billing, teams quietly shorten sequences. Under seat pricing, teams limit who gets a login. Under connected-call billing, the number to watch is conversations, which is the same number sales leaders already care about.

None of these effects is a disaster. But pick the model whose incentives match how you want your team to work.

The hidden line items

Whatever the headline model, ask about:

• Telephony. Included or extra? Per minute or per number?
• Phone numbers. Monthly fees per number?
• Transfers. Are transfer minutes billed, and on whose line?
• Voicemail drops. Billed as a call, as minutes, or not at all?
• Recording and storage. Extra?
• Minimums. Monthly spend, seat minimums, or annual contracts?
• Setup and onboarding fees.
• Overage rates above plan limits.

Some of these can change your real cost more than the headline rate.

Toolkits versus finished workflows

Price comparisons often mix two different kinds of product.

Developer platforms like Bland, Retell, and Vapi are toolkits. They give engineers the building blocks for voice agents: models, voices, telephony connections, and APIs. They are flexible, and their per-minute pricing reflects that. If you have engineers and want to build exactly what you need, they make sense.

A finished workflow like DialSwarm comes with sales features already built: lead upload, call schedules with retries, voicemail rules, warm transfers to a rep's phone, live listen-in and mid-call coaching, call history, opt-in recordings, HubSpot sync, an API and webhooks, and Zapier or Make for speed to lead. You pay for the workflow, not just the minutes.

When comparing, count the cost of building and maintaining what a toolkit leaves to you. Neither path is better in general. They suit different teams.

How DialSwarm prices, and why

We bill connected calls only because we built DialSwarm as salespeople, for salespeople. Reps know most dials do not connect and that persistence is part of the job. Charging for every ring punishes good follow-up. Charging for conversations keeps cost aligned with what actually matters.

Our prices are published, so you do not need a sales call to compare. DialSwarm bills connected calls only, so unanswered dials don't count. You get 10 free calls with no card, and current plans are on the pricing page.

Cheap calls that break the rules are expensive. Whatever you choose, confirm it fits how you must operate: AI voice calls need prior express consent under the FCC's February 2024 ruling (FCC Declaratory Ruling FCC-24-17, accessed Oct 17, 2026). DialSwarm agents say they are AI, call only consented leads, and honor stop requests.

Budgeting month to month

Connected-call pricing makes spend easier to explain to finance. Spend rises when conversations rise, and conversations should line up with transfers and meetings. Review those three numbers together each month. If spend rises but transfers do not, look at your questions and lead sources before you look at the price.

A fair comparison process

1. Shortlist two or three tools that meet your compliance and workflow needs.
2. Run a small test of each on similar, consented leads.
3. Track connected calls, conversations that reached a rep, meetings booked, and total cost including hidden line items.
4. Compute cost per useful conversation for each.
5. Factor in build and maintenance time for toolkits.
6. Decide.

Cost per useful conversation is the number that matters. Per minute and per dial are just inputs.

Questions