Genesys Cloud AI Experience tokens are the unit in which Genesys meters most of the AI features of Genesys Cloud CX. An organisation consumes tokens as it uses the features, the quantity consumed appears on the invoice, and the tokens renew each month and do not carry over to later months.[1] Tokens sit on top of the per-user edition: a qualifying CX licence is a prerequisite, but holding the licence does not mean the feature is included at no cost.[1]
Editions
A Genesys Cloud AI Experience product requires a qualifying licence from the CX 1 to CX 4 lines. The Genesys table marks which licences and licence-plus-add-on combinations qualify for each product, and states that a checkmark means the licence qualifies you to use the product, not that it is included at no additional cost.[1] Examples from the table include:
- Agent Copilot, AI Translate, AI summaries and insights, voice bots, predictive routing, Virtual Agent, Agentic Virtual Agent and Genesys Cloud Copilot qualify on CX 1.[1]
- AI Scoring, digital bots, Predictive Engagement and the third-party messaging channels need a digital-capable licence such as the CX 1 Digital Add-on II, CX 2, CX 2 Digital, CX 3, CX 3 Digital or CX 4.[1]
- WhatsApp messaging and Genesys Cloud Social Listening are not marked for CX 2 but are marked for CX 2 Digital and the CX 3 and CX 4 licences.[1]
On the Genesys pricing page every edition shows AI Studio, virtual agents, native bots, predictive routing and knowledge features as Tokens, while Agent Copilot functions are Included only in CX 4.[7] Genesys states that Agent Copilot is automatically included in CX 4 and that CX 1, CX 2 and CX 3 customers can obtain it through AI tokens.[5] CX 4 also carries a fair use allocation of 30 AI Experience tokens per named user and 39 per concurrent user.[8]
Metrics
The token is the metric: AI Experience token. Each feature has its own raw meter that Genesys converts to tokens. The Resource Center metering table gives the following units:[1]
| Feature | Units per token or tokens required |
|---|---|
| Voice bots | 17 minutes per token; each call is rounded up to the next 15-second increment |
| Digital bots | 51 sessions per token |
| Virtual Agent | 0.5 tokens per interaction |
| Agentic Virtual Agent | 1.2 tokens per interaction |
| Agent Copilot | 40 tokens per named user; 60 per concurrent user |
| AI Scoring | 20 evaluations per token |
| AI Translate | 2 translations per token |
| AI summaries and insights | 50 per token (Supervisor Copilot summaries do not consume tokens when Agent Copilot is enabled) |
| Apple Messages for Business, Facebook Messenger, Instagram Direct Messaging, WhatsApp, X Direct Messaging | 400 inbound or outbound messages per token |
| Genesys Cloud Social | 400 social post ingestions per channel per token |
| Social Post Responses | 400 outbound messages per channel per token |
| Predictive routing | 17 routes per token |
| Speech and Text Analytics | 30 tokens per named user; 45 per concurrent user |
| Genesys Cloud Copilot | 20 AI actions per token; no charge for knowledge queries |
| Predictive Engagement | No charge for token usage |
A Virtual Agent interaction is contained by a single billing ID and ends on events such as an exit or disconnect action, a transfer to ACD, a recognition failure, agent escalation or a 72-hour digital inactivity timeout; transfers between inbound flows and Virtual Agent flows keep the same billing ID.[1]
The revision history shows that the rates change. On 2026-06-01 Genesys reclassified Agentic Virtual Agents powered by Guides as Virtual Agents, added a new Agentic Virtual Agent offering and removed AI Guides from token charges; on 2026-01-13 it raised social post ingestions per token from 200 to 400; and on 2026-07-12 it clarified rounding for voice bots.[1] The page also records that Virtual Agent charging moved from no charge to token-based billing as of 2024-12-15.[1]
Counting / floors
Genesys calculates token use per billing cycle. Tokens used in the cycle are added together and then rounded up, and Genesys does not round up at the individual item level; in its example, 2.5, 5.95 and 60 unrounded tokens total 68.45 and are billed as 69.[3] When an interaction uses several AI resources, such as bot flows, virtual agents and agentic virtual agents, charges follow the highest-tier (highest price) resource used during the interaction.[1]
Free allocation and overage
Each organisation receives a free monthly allocation of 250 tokens if it is a named organisation and 350 if it is a concurrent one.[1] This allocation is per organisation, not per licence.[2] Additional tokens can be purchased, and the pricing hub lists a month-to-month charge of USD 1.00 per token (CAD 1.10, GBP 0.70, EUR 0.80, JPY 120).[3]
The AI token billing article walks through the arithmetic with voice bots. A named organisation’s 250-token allowance equals 4,250 minutes (250 x 17); usage of 15,912 minutes converts to 936 tokens, leaving an overage of 686 tokens, charged at USD 1 each, or USD 686.[2] Genesys bills overage monthly in arrears regardless of whether the contract is monthly or annual.[2] The invoice then spreads the charge across all tokens consumed, including the free ones at a zero rate, so the invoice rate column can show an average (USD 0.74 in the example) rather than the published rate.[2] A reconciliation therefore starts from the usage report in tokens, not from the invoice rate.
Virtualization & partitioning
No server-based partitioning concept applies to tokens. The relevant partition is the Genesys Cloud organisation, because the free allocation is defined at the organisation level, not per user.[4] Whether a company with several organisations receives the allocation for each should be confirmed against the Services Order.
Cloud / BYOL
Tokens apply only to Genesys-operated AI features. Genesys also lists a separate Bring Your Own Technology services model, with rate packages A to E, for third-party text-to-speech, speech-to-text and bot connectors, which are charged per turn or per minute and require the customer to contract separately with the third-party provider.[3] Genesys notes that other AI-powered capabilities such as Genesys Cloud AI Experience Orchestration do not consume AI Experience tokens.[1]
Programs
New customers have access to Genesys Cloud AI and the tokens by default, while existing customers order tokens through AppFoundry or their Genesys representative.[1] Organisations can prepay tokens monthly or annually. Prepaid tokens appear on the invoice as a separate line at the Services Order rate, and overage applies only when usage exceeds both the free allocation and any prepaid commitment.[2] Catalog row: AI Experience token prepay and overage.
Legacy AI bundles
Before tokens, Genesys sold the Genesys Cloud CX AI Experience for Digital and Genesys Cloud CX AI Experience add-on bundles under a per-invocation rate structure, with fixed allocations such as 1,000 digital bot sessions, 20,000 Predictive Engagement events and 1,000 predictive routing interactions per named agent.[6] The pricing hub lists these legacy bundles at USD 40 (digital) and USD 60 (full) per agent on annual prepaid terms and says they continue until the contract is renewed; it also lists legacy Agent Assist SKUs priced per named or concurrent agent until the customer moves to Agent Copilot.[3] Genesys states that it removes the deprecated AI Experience Bundle and deprecated discrete AI products at renewal, so renewal is the moment at which an organisation moves to the token model.[1]
For Facebook Messenger, Instagram and WhatsApp, customers with contracts from before May 2024 that have not added tokens are billed legacy conversation-based rates of USD 0.15 per conversation; adding AI tokens moves them to message-based token rates.[3]
Out of scope
This article does not cover the per-user edition prices, fair use allowances for non-AI resources, Meta pass-through charges for WhatsApp, X integration tier fees, or negotiated token prices. For the free allocation of wallboards, rule executions and cases, see the free allocation program and Genesys Cloud fair use, add-ons and usage billing.