PolyAI pricing

About

PolyAI builds enterprise voice AI assistants — conversational agents that answer phone calls for contact centers and resolve customer requests (account management, authentication, billing & payments, bookings, order management, troubleshooting) across industries like financial services, healthcare, hotels, insurance, retail, telecom, travel, and utilities. The company spun out of the University of Cambridge; co-founder and CEO Nikola Mrksic previously worked on Apple’s Siri, and PolyAI develops its own proprietary voice models rather than wrapping third-party LLMs.

By late 2025 PolyAI reported 100+ enterprise customers and 2,000+ live deployments across 45+ languages — named customers include PG&E, UniCredit, Caesars, Golden Nugget, FedEx, and Marriott. Revenue grew from $8.9M (FY ending Jan 2024) to roughly $15M (FY ending Jan 2025), with projected ARR of $40M+. In December 2025 it raised an $86M Series D at a $750M valuation (co-led by Georgian, Hedosophia, and Khosla Ventures, with NVIDIA’s NVentures and others), pushing total funding past $200M.

For the most current information, visit PolyAI.

Pricing summary : How PolyAI’s pricing model works

PolyAI’s pricing is fully sales-led and gated — there are no published plan tiers or dollar figures. The pricing page makes exactly one concrete commitment about the model: “Ongoing use of the voice agent is priced on a per-minute basis, which includes proactive performance improvements, maintenance and 24/7 support.” Everything else routes the buyer into a “Request a demo” form whose first qualifying question is annual call volume (ranges run from “less than 10,000” up to “more than 1,000,000” calls per year).

So the value metric is minutes of conversation handled by the voice agent, quoted per customer. Per-minute rates vary with the speech-to-text pipeline and multilingual voice configuration used in production, plus call volume, languages, and integration scope. That usage sits inside an annual or multi-year commitment with a minimum spend, and engagements typically begin with a one-time voice-design / solution-design effort to match the agent to the brand and workflows. Every deal — regardless of size — bundles a 99.9% uptime SLA, 24/7/365 support, security/compliance certifications, monitoring, and ongoing upgrades.

What makes this different: PolyAI deliberately keeps prices off the page and leads with a volume question, because the per-minute rate only makes sense once call volume, languages, and integrations are scoped. The meter (minutes) maps directly to a contact center’s existing telephony economics — buyers already think in call-minutes — but the lack of any public rate means cost clarity arrives late in the sales cycle.

Pricing by product

Sales motions across products: 100% sales-led. There is no self-serve signup, no free tier, and no published rate card — access is via a demo request that captures annual call volume first. Pricing is scoped in a solution-design conversation and lands as an annual/multi-year commitment.

Hidden costs : What PolyAI users actually pay

Because PolyAI publishes no rate, the real cost is whatever the quote works out to — and several dimensions move it. The figures below are third-party-reported and indicative only; PolyAI does not confirm contract sizes, setup fees, or rates.

Other things to budget for: PolyAI does not disclose minimum contract length, setup fees, or per-minute rates publicly, so total cost of ownership is only knowable after a scoping workshop. Heavy multilingual configurations and more complex speech pipelines reportedly raise the per-minute rate, and deeper backend integrations expand the implementation cost.

Want to estimate your own PolyAI bill? Use the PolyAI pricing calculator to model your costs based on usage patterns.

Pricing evolution : PolyAI pricing history and changes

Cadence

Tracked range: 2026–present. PolyAI has never published plan prices, so there is no public price-history to track via Wayback — the page has consistently been a gated demo-request flow describing per-minute billing.

Notable changes

• 2026-06-09 — Pricing page confirms a single sales-led, per-minute model gated behind a “Request a demo” form (qualifying on annual call volume), with a 99.9% SLA, 24/7 support, security, monitoring, and upgrades bundled into every engagement. No dollar figures are published.

What’s unique : PolyAI’s distinctive pricing mechanics

1. Per-minute, not per-resolution or per-seat. PolyAI meters the actual minutes of conversation its voice agent handles — a metric that maps cleanly onto a contact center’s existing telephony cost model. Many newer voice-AI vendors quote per-resolution or per-call; PolyAI’s per-minute basis ties price to the work performed.

2. Volume is the first question, not the last. The demo form leads with annual call volume (up to “more than 1,000,000”) before anything else. Pricing is genuinely volume-derived, so the rate only exists after the buyer is bucketed — there is no list price to anchor on.

3. Everything-included SLA bundle. Rather than upselling support, security, and uptime as add-ons, PolyAI folds a 99.9% SLA, 24/7/365 support, compliance, monitoring, and upgrades into the per-minute price for all customers — positioning it as a managed enterprise service, not a self-serve API.

Strengths & weaknesses

Billing UX : PolyAI billing controls and transparency

• Billing controls — Enterprise-contract based: there is no self-serve dashboard for plan changes. Spend is governed by the negotiated annual/multi-year commitment and per-minute rate, adjusted through the account team rather than in-app.
• Usage visibility — PolyAI offers Data and Insights tooling for conversation analytics, but it does not publish a public cost calculator or per-minute rate, so forward cost visibility for prospects is limited until a quote is issued.
• Payment options — Invoiced enterprise billing under a custom contract; the demo form collects company, role, location, and annual call volume to scope a quote. No card-based self-serve checkout.

Strategic wins : Why PolyAI’s pricing decisions worked

1. Metering on minutes the buyer already counts

By pricing per minute of conversation, PolyAI speaks the contact center’s native language — call-minutes are already a budgeted line item, so the meter feels familiar and the ROI case (deflected agent-minutes) is easy to build. See choosing the right usage metric.

2. Bundling the managed-service promise into one rate

Folding the 99.9% SLA, 24/7 support, security, and upgrades into the per-minute price positions PolyAI as a dependable managed partner for mission-critical phone lines — not a DIY API. For enterprises automating their front door, that reliability framing justifies a premium and a committed contract. Related: how AI companies structure pricing.

3. Volume-first qualification

Leading the demo flow with annual call volume lets sales segment and price around the one variable that most determines value and cost. It keeps PolyAI focused on high-volume enterprise accounts where per-minute economics work. See outcome-based pricing trends.

Areas to improve : Gaps in PolyAI’s pricing approach

1. Zero public price anchor

Even sales-led vendors increasingly publish a starting per-minute rate or a “from $X” anchor. PolyAI publishes none, forcing prospects to rely on third-party review sites for rough figures — a transparency gap that slows evaluation. See bill shock and cost unpredictability.

2. Opaque rate drivers

The per-minute rate reportedly varies with STT pipeline, voice configuration, and languages, but PolyAI does not explain how. Buyers can’t model how multilingual or complex deployments will change their bill until a quote arrives.

3. High floor excludes the mid-market funnel

A reported entry around $150,000/year plus implementation fees keeps PolyAI firmly upmarket. That protects margins but cedes the growing mid-market voice-AI segment to self-serve, per-minute competitors that publish rates and offer trials.

Key takeaways

1. PolyAI is sales-led and gated. No plan prices are published; the only concrete public fact is that ongoing voice-agent use is billed per-minute.
2. The value metric is conversation minutes. Rates vary with volume, languages, and STT/voice configuration, scoped per customer in a solution-design conversation.
3. Usage sits inside a committed contract. Engagements are annual/multi-year with a reported ~$150K minimum (third-party), plus one-time voice-design fees.
4. Reliability is bundled, not sold. A 99.9% SLA, 24/7 support, security, monitoring, and upgrades are included in every deal — the managed-service framing.
5. It’s a deliberate enterprise posture. A Cambridge spin-out with proprietary models, $750M valuation, and 100+ enterprise customers, PolyAI prices for high-volume contact centers, not SMB self-serve.

UBP implications

1. Per-minute is a clean meter for voice AI. Tying price to minutes handled maps to the buyer’s existing telephony economics and to the value delivered (deflected human-agent time). See usage-based pricing strategy.
2. Usage + commitment is the enterprise default. Wrapping per-minute usage in an annual minimum gives the vendor predictable revenue while still scaling with the customer’s volume — a common enterprise UBP pattern.
3. Gated pricing trades clarity for control. Keeping rates private lets PolyAI price by value and segment, but raises the evaluation burden and pushes price-sensitive buyers toward transparent, self-serve competitors.

Sources

• PolyAI pricing page (live capture, accessed 2026-06-09)
• PolyAI raises $86M Series D at $750M valuation — SiliconANGLE (accessed 2026-06-09)
• Voice Assistant Builder PolyAI Closes $50M Series C — PR Newswire (accessed 2026-06-09)
• PolyAI Pricing 2026 Breakdown — Nurix (third-party pricing estimates; accessed 2026-06-09)
• PolyAI Pricing: Plans, Costs, and Best Alternatives — CloudTalk (third-party pricing estimates; accessed 2026-06-09)
• PolyAI — Crunchbase company profile (accessed 2026-06-09)

Bottom line

PolyAI is a Cambridge spin-out building enterprise voice AI for contact centers, valued at $750M after an $86M December 2025 Series D and serving 100+ enterprises (PG&E, UniCredit, Caesars, FedEx, Marriott) with 2,000+ deployments in 45+ languages. Its pricing is fully sales-led and gated: the only public fact is that ongoing voice-agent use is billed per-minute, with rates scoped per customer by volume, languages, and configuration and wrapped in an annual/multi-year commitment (reported ~$150K minimum plus one-time voice-design fees — third-party figures). A 99.9% SLA, 24/7 support, security, monitoring, and upgrades are bundled into every deal. Browse the pricing blueprint for more fully-researched company profiles.

Want to compare PolyAI against other voice and contact-center AI companies? Browse the pricing blueprint.

Pricing timeline : Major events on a vertical axis

Each milestone below corresponds to a public pricing change, product launch, or material adjustment. Major events use a filled marker; minor adjustments use a faded one.

Per-minute, sales-led pricing (gated)

Jun 2026

PolyAI's pricing page presents a single sales-led motion: ongoing voice-agent use is billed on a per-minute basis (including performance improvements, maintenance, and 24/7 support). No plan tiers or prices are published — access is via a 'Request a demo' form that captures annual call volume. All engagements include a 99.9% uptime SLA, 24/7/365 support, security/compliance, monitoring, and upgrades. Third-party sources report a ~$150K annual minimum plus one-time voice-design fees.