Why Advertised Per-Minute Rates Mislead You

That "$0.05 per minute" rate on a vendor's pricing page is rarely the number that lands on your invoice. When a platform advertises a rate that low, it may cover only one cost layer, leaving you to pay separately for everything else that makes the call actually work, according to a detailed pricing breakdown.

The problem is structural. Every AI call carries four distinct cost layers: telephony ($0.01–$0.03/min), speech-to-text ($0.01–$0.02/min), the LLM ($0.01–$0.04/min), and text-to-speech ($0.03–$0.10/min). Platforms using "Bring Your Own Key" models advertise low base prices but require you to add roughly $0.06–$0.19/min in provider costs before the platform's own margin even applies, per a 2026 pricing analysis. A rate that looks like a bargain can quietly double.

Beyond the four layers, hidden billing events inflate the total further. Buyer reports identify voicemail billing, rounding rules, and charges for silence as consistent sources of unexpected spend — and overage minutes can run two to three times above the bundled effective rate, according to pricing guides. For outbound patient campaigns, voicemail detection and unanswered calls matter most, since those attempts may still be billed as connected time.

Watch for these common hidden charges:

• Billed silence and hold time, where a call that never really starts still counts against your minutes.
• Rounding rules that bill a 40-second call as a full minute, compounding across thousands of calls.
• Voicemail detection fees on outbound campaigns, charged even when no patient conversation occurs.
• Add-ons for model choice, voice choice, knowledge base access, concurrency, or telephony layered on top of the base rate.

The consequences of judging cost on the sticker rate are measurable. At 30% containment, the cost per call actually resolved by AI reaches $1.30 — more than double the $0.60 shown on the vendor slide — because failed attempts get folded into the math, per vendor-neutral cost analysis. Teams that judge cost per call on wages or headline rates alone routinely understate the real figure.

The fix is to ask one question before signing anything: what are my exact total costs for a given number of minutes, assuming the AI model, voice, transcription, and telephony are completely included? Connected minutes should mean conversation time, not everything the vendor decides to bill. My AI Call Center quotes campaigns this way — the full number, including connected minutes at a rate agreed before launch, is known before anything goes live.

Cost per resolved patient interaction, not advertised per-minute price, is the only number that tells you whether AI calling actually saves money.

The Cost Per Patient Formula, Step by Step

Most clinics can quote their per-minute AI rate instantly — but very few can tell you what a single resolved patient interaction actually costs. The difference between those two numbers is where budgets quietly fall apart.

The core formula is simple: Cost per resolved patient interaction = Total operating costs ÷ Number of successfully resolved patient contacts, measured over the same period. As cost-per-contact analysis puts it, the metric is your total contact-center operating cost divided by contacts handled in that same window — and the definition of "contact" must stay consistent or comparisons break.

The numerator is where accuracy lives. A defensible figure includes far more than the advertised rate:

• Platform and management fees, plus one-time setup and integrations ($1,000–$5,000 is typical, per pricing research)
• Telephony, speech-to-text, LLM, and text-to-speech costs — BYOK platforms add $0.06–$0.19/min on top of base fees
• Compliance overhead (HIPAA and similar requirements can add $100–$500/month)
• Failed attempts and transfers — the AI minutes spent on calls that end with a human anyway

Because per-minute rates hide these layers, use the companion formula: True Cost/Min = (Monthly fees + provider costs + telephony) ÷ total minutes used. Pricing breakdowns show that a "$0.05/min" headline can easily double once all four cost layers are counted. Confirm which events are billable, too — some vendors charge for silence, hold time, or voicemail detection.

Worked example: a contact center spending $180,000 monthly across 30,000 calls has a cost per contact of $6.00. That sits near the industry benchmark of roughly $6 per general call, with complex healthcare support often exceeding $10.

Now apply the resolution lens. Containment analysis shows that at 30% containment, cost per call the AI actually resolves is $1.30 — more than double the $0.60 on the vendor slide — because failed attempts stay in the math. A managed service like My AI Call Center quotes the full campaign — setup, management, and per-connected-minute rate — before launch, so both inputs of the formula are known in advance, not discovered on the first invoice.

The number that matters is cost per resolved interaction, not the rate card. Everything else is arithmetic waiting for honest inputs.

Connected Minutes and Containment: The Two Numbers That Change Everything

The real cost of a patient interaction isn’t what you see in the per-minute rate—it’s what you pay for each call that actually gets resolved. Many vendors advertise rates as low as 9¢ per connected minute, but that figure only tells part of the story. When calls fail to resolve and escalate to human agents, the true cost per successfully resolved interaction rises sharply. This is where containment rate becomes the decisive factor in determining whether AI delivers real savings.

At My AI Call Center, we bill strictly for connected minutes—only the time when a live person is engaged with the AI voice. This means no charges for ringing, silence, hold time, or voicemail detection. However, even with transparent billing, low containment inflates the effective cost per resolution. Research shows that at just 30% containment, the cost per call actually resolved by AI exceeds $1.30—more than double the vendor-quoted AI-only cost of $0.60 per resolved call. This happens because the AI spends minutes on failed attempts that still require human follow-up, and those costs are spread across fewer successful outcomes.

Containment rate directly impacts break-even economics. Against onshore human agents, AI becomes cost-effective at approximately 11% containment. For offshore teams, the threshold rises to 43% containment due to lower baseline labor costs. Below these levels, AI increases your cost per resolved interaction; above them, savings compound rapidly. For example, at 70% containment, AI can save over $92,000 per month compared to an onshore team handling the same volume.

• Cost per resolved patient interaction = Total operating costs / Number of successfully resolved patient contacts
• At 30% containment, AI cost per resolved call exceeds $1.30—more than double the quoted $0.60 AI-only figure
• Break-even containment: 11% for onshore staff, 43% for offshore staff

Focusing on cost per successfully resolved interaction—not cost per minute—reveals the true economics of AI in patient outreach. It accounts for failed attempts, supervision, integrations, and all operational layers that vendors often exclude from headline pricing. This shift in measurement is essential for clinics, franchises, and multi-location businesses seeking to run more useful calls without building a bigger call center. It’s the only way to know whether your AI investment is truly reducing cost per patient—or just moving expenses around.

How Clinics Should Calculate Healthcare ROI

Healthcare clinics looking to measure the true value of AI-powered calling must move beyond simple per-minute pricing to calculate ROI using a comprehensive framework. The healthcare ROI equation—ROI (%) = [(Total financial benefit − Total investment) / Total investment] × 100—requires capturing all sources of financial benefit: labor avoided, QA cost avoided, and revenue recovered from improved appointment conversion and reduced no-shows. This approach ensures clinics account for both cost savings and revenue gains, not just operational efficiency.

A worked example from a 150-agent healthcare contact center illustrates this in practice. The center achieved $728,000 in labor savings by avoiding 14 planned hires at a fully loaded cost of $52,000 each, plus $174,000 in QA cost avoidance from redeploying three analyst roles. Revenue impact totaled $693,000, combining $198,000 from a two-point conversion gain on 60,000 scheduling calls and $495,000 from a three-point reduction in no-shows across 100,000 scheduled visits. With total financial benefit at $1,595,000 and total investment of $400,000 ($260,000 in year-one licensing and $140,000 in implementation), the net ROI was $1,195,000, resulting in a 299% return—nearly 4x the initial investment—with payback achieved by month five.

To ensure accuracy, clinics should lock a 3–6 month baseline of historical data before launching any AI calling initiative. This baseline must capture cost, labor, call-center performance, patient experience, quality, and revenue metrics to provide a defensible comparison post-implementation. Without this foundation, ROI calculations risk overstating benefits or missing hidden costs like failed AI attempts that escalate to human agents. As research shows, at just 30% containment, the cost per call resolved by AI can exceed $1.30—more than double the vendor-quoted AI-only cost—due to inclusion of failed attempts.

For clinics using managed services like My AI Call Center, this means focusing on connected minutes billed only for actual engagement, not silence or voicemail, while tracking outcomes like confirmed appointments, qualified leads, and reduced no-shows. The goal is not just to lower cost per contact but to increase the proportion of successfully resolved patient interactions—whether that’s a confirmed visit, a completed survey, or a retained patient—thereby improving both efficiency and revenue capture. By grounding ROI in real-world outcomes and layered cost structures, clinics can make informed decisions that scale value without overbuilding internal capacity.

How My AI Call Center Makes the Math Simple

The hardest part of any cost-per-patient calculation isn't the division — it's knowing what number to put on top. Most clinics never get a clean numerator because vendors bury platform fees, provider charges, and overage rates inside a single invoice. Research on AI voice pricing shows why this matters: a platform advertising "$0.05/min" may cover only one cost layer, leaving you to pay separately for telephony, speech-to-text, LLM, and text-to-speech, which can add $0.06–$0.19 per minute before the platform's margin even applies (SigmaMind's cost breakdown, Famulor's pricing analysis).

My AI Call Center was built to remove that ambiguity from the start. Calling starts at 9¢ per connected minute, tiered by volume, and the rate is agreed before launch — it does not move mid-campaign. Most campaigns add a one-time setup and a flat monthly management fee, both quoted upfront. There are no per-seat charges, no platform bill, and no minimums you didn't choose, so your numerator is fully known before you approve anything.

The second half of the formula — the denominator — is where most AI cost calculations quietly fall apart. As one analysis notes, cost per resolved call is the only number that tells you whether AI saves money, because it counts the AI minutes spent on calls that end up with a human anyway (Zenthos's AI call center cost analysis). At 30% containment, the cost per call the AI actually resolves can exceed $1.30 — more than double the vendor-quoted figure (Zenthos).

That's why every campaign ends with a disposition-coded outcome report. Each contact comes back tagged — confirmed, qualified, renewed, opted out, or no answer — with per-call notes and follow-up requests routed into your CRM. You can compute cost per confirmed appointment or cost per renewed patient from what actually happened, not from a projection. It's the same principle healthcare analysts recommend: a defensible ROI calculation starts with a locked baseline of 3 to 6 months of real data (The Level AI's healthcare ROI guide).

Here's what makes the math simple:

• A locked rate — connected-minute pricing from 9¢, agreed before launch, never moved mid-campaign.
• Known fixed costs — one-time setup and flat monthly management fee, both quoted before you approve.
• Real denominators — disposition codes let you divide by confirmed appointments or renewals, not by attempts.

The first campaign review is free, and it establishes your baseline: what the calls cost, what they confirmed, and what each outcome actually cost you. No invented numbers — just the arithmetic, done on your data.