Voice AI cost calculator for Genesys Cloud: native and third-party compared

This calculator is a guide to how each pricing model behaves, not a quotation, a recommendation to buy, or a statement of what any organisation pays. It uses only the list prices each vendor publishes, checked on the date shown beside the calculator, and it applies them to the volumes you enter in a deliberately simple model. Real costs depend on your contract, any discount or prepaid commitment, your currency and taxes, how calls actually flow, and costs the calculator leaves out entirely. Prices change, sometimes with little notice, and vendors' own pages are always the authority.

QVCCS is a Genesys Cloud CX consultancy and a Genesys referral partner, not a reseller. We do not sell licences, tokens or third-party AI, and we receive no commission or payment from any vendor shown here. We built this because buyers asked us the same question again and again, and the honest answer depends on their own numbers. If any figure here is wrong or out of date, we want to know and will correct it.

• Excluded: telephony and carrier minutes, which every route pays for alike
• Excluded: design, build, testing, tuning and support, for any route
• Excluded: speech output for Amazon Lex, and the language model for ElevenLabs unless you enter it
• Excluded: enterprise plans, volume discounts, prepaid tokens and minimum commitments
• Third-party prices published only in US dollars are converted at the European Central Bank reference rate shown

Enter your monthly calls, how many you would offer to AI and how many it should resolve, how long each call spends with the AI, and your agent handle time. Tell it about your Genesys organisation, because other AI features share the same free token allowance. Add your fully loaded cost per agent hour to see each route all-in against handling every call with agents. Everything runs in your browser: nothing you type is sent anywhere or stored.

For Genesys's native options, the calculator converts usage into AI Experience tokens at the published rates, then charges only the tokens that the AI adds above your free allowance once other token users such as Agent Copilot are counted. Voice bot minutes are rounded up to fifteen seconds per call before conversion. For third-party routes, it multiplies AI minutes by the vendor's published rate and adds the Genesys platform charge for that connection: Rate C for the native Amazon Lex V2 and Google Dialogflow integrations, Rate E for routes through Audio Connector.

Some vendors bill on less than the whole call. Google bills only the seconds its agent is listening or speaking, and Amazon Lex counts listening time in fifteen-second intervals but not its own responses. The calculator charges the full AI minutes for both, so their figures are upper estimates. Calls the AI offers but does not resolve are charged for the AI and, if you entered an agent cost, for the agent who then handles them.

1. AI calls = calls a month x share offered to AI; resolved = AI calls x share resolved
2. Native tokens: voice bots per minute, Virtual Agent and agentic agent per interaction
3. Token charge = tokens added above the shared free allowance x list price per token
4. Third-party charge = AI minutes x vendor rate, plus the Genesys platform rate
5. All-in = AI charge + agent time for every call the AI does not resolve

These are the figures behind the calculator, each from the vendor's own pricing page on the date shown. Genesys publishes its prices in pounds, dollars and euros; Google, Amazon, Deepgram and ElevenLabs publish in dollars, so their pound and euro figures are conversions and marked as such. The second table shows Genesys's token consumption rates and allowances, which are not prices but decide how many tokens are charged.

Published by each vendor and checked on 14 October 2026. These are list prices, before any discount, tax or contract terms, and they change; the sources at the end of this page are authoritative. QVCCS does not sell licences. Prices published only in US dollars are converted at the European Central Bank reference rates of 2026-10-13 (1 EUR = 1.1188 USD = 0.8446 GBP).

Price is only one column of a sound decision. A native Genesys agent is built, tested, reported and quality-managed inside the platform you already run, and its conversations sit in the same analytics and recordings as your agents' calls. A third-party route can offer a particular voice, language model or capability, but it adds a second vendor, a second contract, a second place where customer data is processed and a hand-off that has to be designed and tested. Where the language model runs, and under whose terms, is as much part of the decision as the price per minute.

The ranking also moves with your inputs. Short, structured calls favour scripted bot flows, which are metered by the minute; long conversations that need reasoning favour charges per interaction. Containment matters more than any rate, because every unresolved call pays twice. Treat the results as a starting point for a conversation with each vendor, and test any shortlisted route with real calls before you commit.

We chose routes a Genesys Cloud CX buyer would realistically shortlist and whose prices are published. Managed services that quote individually are listed without a figure rather than estimated. We will add routes as vendors publish prices and as Genesys extends its integrations.