Voice AI pricing models: per-minute, per-resolution, and platform compared

Per-minute pricing

Per-minute is the simplest model and the easiest to compare on a rate card. It transfers no containment risk to the vendor — the buyer pays for every minute the AI is on the line regardless of outcome.

Typical 2026 enterprise per-minute pricing for a fully managed voice AI stack lands in the range of $0.10 to $0.35 per minute, with model choice and telephony bundling the largest swings. Self-hosted or build-it-yourself stacks routinely run lower per minute and higher per total cost of ownership.

Per-resolution pricing

Per-resolution pricing charges only for calls the AI resolves end-to-end, with a written definition of 'resolved' attached to the contract. It transfers containment risk to the vendor — which both aligns incentives and forces the contract to define resolution before signing.

Typical 2026 enterprise per-resolution pricing lands in the $1.00 to $4.00 range, depending on call complexity and what 'resolved' includes. The model is most defensible when the buyer is early in its voice-AI journey and the containment rate is genuinely uncertain.

Platform pricing

Platform pricing decouples cost from per-call volume. The buyer pays a fixed annual platform fee plus a usage component that is small relative to the platform fee. The model favours buyers running at predictable, high volume and tends to disadvantage low-volume or seasonal deployments.

Cost per resolved call — the only fair comparison

Convert each model to cost per resolved call using your modelled containment rate, then stress-test at 0.5x and 2x that rate. Three things almost always come out of that exercise: per-minute looks best at the assumed containment and worst at half of it; per-resolution is the inverse; platform pricing wins decisively at high volume and loses badly at low volume. The right answer is almost never the headline.

Hidden costs that move the answer

Five line items move the comparison materially and are almost never on the rate card: telephony pass-through, model API charges if not bundled, integration / connector fees, professional services for the operating model, and the change order for any custom voice or guardrail work. A 30 to 60% uplift on the headline number is normal once these are included.

A vendor-proposal teardown, line by line

Most enterprise voice AI proposals collapse onto a single per-minute or per-resolution headline. The defensible read is to expand that headline into the eight cost lines that actually show up on the first invoice and the four that show up later. Doing this once across three proposals usually re-orders the shortlist.

• Headline rate — per-minute, per-resolution, or platform; check what is included and what is bundled separately
• Telephony pass-through — SIP termination, inbound DIDs, geographic surcharges; often a 10–25% uplift
• Model API charges — sometimes bundled into the headline, often passed through with a margin; ask whether you can bring your own model contract
• Speech-to-text and text-to-speech — usually bundled, sometimes itemised; voice cloning typically extra
• Recording and storage — retention beyond a default window is almost always a chargeable line
• Integration connectors — generic connectors usually free, named systems of record sometimes a per-connector fee
• Professional services — implementation, conversation design, prompt engineering; budgeted at 15–35% of year-one platform spend in practice
• Change orders — the line item that catches everybody; tariff for net-new intents, prompt redesigns, and custom voices
• Volume floor — the minimum committed minutes or resolutions; rarely advertised, almost always in the contract
• Annual uplift — most contracts now embed a 3–7% annual price uplift; negotiate it down at signature, not at renewal
• Egress and audit access — exporting transcripts or recordings in bulk for QA or audit sometimes carries an API charge
• Exit fee — data return and provable destruction at contract end; non-zero, sometimes punitive

A worked GBP example: 600,000 calls a year

Take a UK contact centre with 600,000 inbound calls a year, average pre-AI handle time of 6 minutes, and a fully loaded agent cost of £38 an hour. Pre-AI direct handling cost is roughly £2.28 million.

Run the same volume against three pricing models at 35% measured net containment. Per-minute at £0.14/minute lands at platform cost of ~£840k (assuming 10 minutes of AI exposure per call across contained and pre-transfer). Per-resolution at £1.80/resolved-call lands at ~£378k for the 210,000 contained calls plus a per-minute floor on escalated calls (~£294k) — total ~£672k. Platform pricing at £750k flat plus £0.04/minute overage lands at ~£990k.

On these inputs per-resolution is cheapest at the modelled containment. Drop containment to 18% and per-minute wins on cost but per-resolution still wins on risk transfer. Push containment to 50% and platform pricing pulls ahead. The right answer changes with the containment assumption; pretending it doesn't is the most common pricing-model mistake.

Hybrid models that actually balance risk

Three hybrid patterns now appear in most enterprise proposals; each shifts risk in a different direction. Choose the one that matches what you actually need protected.

• Per-resolution with a per-minute floor — protects the vendor against you running a low-containment deployment forever; appropriate when both sides believe containment will land above the floor
• Platform with per-resolution overage — buyer absorbs base risk in exchange for capped marginal cost above forecast; appropriate at high volume with low variance
• Per-minute with a containment SLA — vendor refunds a percentage of fees if measured net containment drops below an agreed band; rare, but the most aligned of the three

The five negotiation points worth the most

Across the proposals practitioners see, the same five negotiation points consistently move year-one spend by 15–30%. Each is easier to negotiate before signature than after.

• Annual uplift cap — anchored to a published index (CPI), not to vendor list
• Definition of 'resolved' — written into the contract, with worked examples and a stated re-contact window
• Volume floor — proportional to your forecast confidence, not the vendor's pipeline target
• Professional services — fixed-price phases against named milestones, not time-and-materials
• Exit terms — data export format, timeline, and proof-of-destruction; the cheapest leverage you will negotiate