--- schema_version: "1" id: 87258053b92e2d5e482b846d12bb860b737600f0e2c55b15334b7ace8884e533 title: Agentforce and Generative AI Usage and Billing collection: admin version: current locale: en-us portal_url: https://docs.salesforce.com/admin/current/en-us/ai/264-0-0/generative_ai_usage.md canonical_url: https://help.salesforce.com/s/articleView?id=ai.generative_ai_usage.htm&release=264.0.0&type=5 source_site: help.salesforce.com content_type: documentation content_hash: efeb7699facd91f669526baead0538d1081d25ade98849c02bbf7fdb0c3dc213 description: Salesforce provides several options for how you pay for AI usage. products: - Agentforce - Cross Cloud release: "264" --- # Agentforce and Generative AI Usage and Billing Salesforce provides several options for how you pay for AI usage. Billing for AI interactions depends on several factors, including the environment where the activity occurs and the nature of the interaction. - The pricing model determines the framework for pricing and metering rules. - AI lifecycle phase determines whether an activity is metered based on its stage in development. ## Pricing Models Salesforce offers three pricing models for AI usage. - Consumption-Based Pricing: You pay for what you use based on foundational AI metrics, such as the number of prompts or actions. - Hybrid Pricing (Per User/Per Month Licensing + Consumption): This model combines fixed user licensing with usage-based consumption to offer price certainty per user for many AI scenarios. - Per-user licensing: Provides unmetered usage for specific features within a defined scope. - License-Specific Metering: Grants unmetered access to features only available through specialized licensing. - Metering Conditions: Triggers Flex Credit consumption only when license-specific conditions aren’t met. See [Metering for Agentforce and Generative AI Usage](https://help.salesforce.com/s/articleView?id=ai.generative_ai_usage_metering_logic.htm&release=264.0.0&type=5 "Generative AI usage for Agentforce features is typically metered and consumes credits.In some cases, when required conditions are met, usage can be unmetered and doesn’t consume credits."). - Business-Metrics-Based Pricing: Under this model, you pay based on high-level business outcomes rather than direct technical consumption. Examples include Agentforce Voice Minutes and characters translated. ## How AI Usage Is Metered and Credits Consumed AI usage falls into two main categories that dictate the specific unit of consumption. - Agentic usage: Agentic usage involves autonomous agents that can reason, perform multi-step tasks, and act independently. The unit of consumption for this type is based on the number of actions triggered. - Prompt-Based (Embedded) AI Usage: These features assist users with specific tasks—such as summary generation or email drafting—without acting autonomously. Consumption is based on the number of prompts, which include direct requests to Large Language Models (LLMs). The system records AI usage in its native units, such as actions or prompts, before converting them into Flex Credits. This process follows three steps: - Usage Type Categorization: These units are mapped to specific usage types based on the specific feature or agent used. - Measurement: The system counts the raw units of consumption based on the usage type (for example, a single action or a single prompt). - Credit calculation: The usage types are converted into Flex Credits by using a unit price or multiplier defined in the [Agentforce Rate Card](https://www.salesforce.com/ap/agentforce/rates/). This multiplier allows different AI services to bill at varying rates against a unified credit pool. See these topics to learn more about the usage types, and their unit of measurement. - [Flex Credits](https://help.salesforce.com/s/articleView?id=ai.usage_flex_credits.htm&release=264.0.0&type=5 "Using Agentforce and embedded AI features can impact credit consumption.") - [Conversations](https://help.salesforce.com/s/articleView?id=ai.generative_ai_usage_conversations.htm&release=264.0.0&type=5 "Usage of certain Agentforce features impact credit consumption against subtypes of the Agentforce usage type on the Conversations consumption card. It consists of three subtypes: ASA Messaging, Sales Coach, and SDR.") - [Einstein Requests](https://help.salesforce.com/s/articleView?id=ai.generative_ai_usage_einstein_requests.htm&release=264.0.0&type=5 "Usage of certain generative AI features impacts credit consumption against subtypes of the Einstein Requests billable usage type. Usage is calculated based on the number of calls to the LLM gateway. Einstein Request usage types apply to embedded generative AI features such as Service Replies that involve direct requests to the LLM for specific tasks.") ![Note](https://sf-zdocs-cdn-prod.zoominsoftware.com/tdta-ai-generative_ai-264-0-0-production-enus/f8ef6498-d7b2-44e3-9b34-ffd5a850aae9/images/icon_note.png) Note To monitor your metered usage access your org’s [Digital Wallet](https://help.salesforce.com/s/articleView?id=xcloud.wallet_about.htm&type=5&language=en_US). ## Metering in the AI Lifecycle The billability of an AI interaction also depends on the stage of the development lifecycle. - Design and Development: Agentforce Builder features for creating and building your agents aren't metered and don't consume credits. However, Salesforce meters usage when you preview an agent in chat or voice. - Testing and Validation: Testing through Agentforce Builder, Agentforce Grid, Testing Center, and Sandbox are metered to account for required compute resources. The type of testing includes the preview options for voice and Chat or running batch tests. - Production: All live interactions in a production environment based on the consumption model: autonomous agent workflows and user-triggered prompts. ![Note](https://sf-zdocs-cdn-prod.zoominsoftware.com/tdta-ai-generative_ai-264-0-0-production-enus/f8ef6498-d7b2-44e3-9b34-ffd5a850aae9/images/icon_note.png) Note Agentforce Grid is commonly used for testing, but its applicability extends across all three phases of the AI lifecycle. Importantly, Agentforce Grid usage is metered regardless of the AI lifecycle phase in which it’s used.