Anti Money Laundering AI pricing


AML AI pricing is based on two factors:

  1. The number of a bank's customers (referred to as parties) a Google Cloud customer uses AML AI for, billed daily
  2. The amount of experimentation a Google Cloud customer uses for training and tuning models against their datasets

Production usage for risk scoring

SKU Tier Price Level1
Risk scoring for Retail Banking usage NA $X.XX per retail party per year
(prorated to usage)
Risk scoring for Commercial Banking usage Small commercial
(<500 transactions / month)
$X.XX per small commercial party per year
(prorated to usage)
Large commercial
(>=500 transactions / month)
$X.XX per large commercial party per year
(prorated to usage)

Production usage is priced by the number of the parties (bank's customers) which are scored for risk.

  • Retail parties are individuals who use banking services for personal usage whereas commercial parties are companies or individuals with accounts used for business purposes. Different model engines and data schemas are used to score for the risk of money laundering in each.
  • Commercial parties are divided into small and large companies based on the average number of monthly transactions for the party over the preceding 365 days.
  • Parties must be registered with the service to obtain predictions. Parties without prediction intent can be removed at any point, but a minimum registration period of 45 days applies to parties with prediction intent, after which the applicable party may be deregistered under the service. No registration is required for training, tuning, or backtesting.
  • Parties are registered per AML AI instance using the instances.importRegisteredParties method. Parties registered in one instance are registered in other instances and will remain registered for a minimum of 45 days if predicted on, before they can be removed from the registry. Billing occurs for each instance separately for the period the customer is registered for.
  • The lists of currently registered parties can be retrieved using the instances.exportRegisteredParties method.
  • You can register both retail and commercial customers on the same instance.

Training and tuning of models

SKU Price Level
Training $X.XXXX per party in dataset
Tuning $X.XXXX per party in dataset

Training and tuning is priced based on the number of parties in the dataset used to train a model or tune an engine. AML AI does training when you create a Model resource and tuning when you create an Engine Config resource.

1 Annual pricing displayed for convenience, all prices are prorated to the period a party was registered for.

Request a custom quote

With Google Cloud's pay-as-you-go pricing, you only pay for the services you use. Connect with our sales team to get a custom quote for your organization.
Contact sales