Pass-through pricing
Pass-through pricing is a model in which a platform charges for its own software or orchestration while billing component-provider costs separately. Those external costs may appear on the platform’s invoice at cost, or the customer may pay the providers directly through its own accounts.
In a voice AI service, the platform fee may cover call logic, workflows, integrations, monitoring, and administration. Speech recognition, language-model processing, voice synthesis, and telephony can then be passed through as separate usage charges. Each component may use a different billing unit, increment, currency, or pricing tier.
Because of that separation, the platform’s pricing page may not reveal the complete operating cost. Calculating a total requires the rates for every selected provider, the amount of each resource consumed, and the platform’s own charges. Call duration alone may be insufficient because model tokens, generated audio, transcription time, carrier destinations, or connected services can vary independently.
Some platforms use a bring-your-own-keys model. The customer supplies credentials for supported AI or voice providers, and those providers bill the customer’s account directly. This arrangement can provide direct access to provider pricing, quotas, logs, and negotiated terms. It also creates more accounts to configure and monitor, and a provider-side limit or billing problem can interrupt the larger call workflow.
“At cost” should be clarified rather than assumed. A platform may pass through the published provider charge without markup, apply a processing or administration fee, or use its own contracted rate as the billing basis. Taxes, currency conversion, minimum commitments, and provider-specific surcharges may also affect the final amount even when the underlying usage is not marked up.
Buyers should ask which components are passed through, which are included in the platform fee, and whether any margin or handling charge is added. They should identify who selects the provider and model, who receives provider discounts, how rate changes are communicated, and whether usage can be reconciled between the platform and provider invoices.
Support ownership is equally important. When a call fails because an external speech, model, voice, or carrier service is unavailable, the customer needs to know whether the platform investigates the full path or expects the customer to work with the provider. Pass-through pricing can make costs more transparent and configurable, but it can also divide billing and operational responsibility across several parties. A meaningful comparison should include both the expected total cost and the work required to manage the component relationships.