Overview

Pay-i provides tools to track, control, and understand costs, performance issues, and failures arising from GenAI consumption. GenAI introduces many new concepts, and so does GenAI FinOps. Therefore, this section describes the high-level terminology used by all Pay-i workflows, some of which may be new to you. To help with contextualizing the concepts described below, we will use the following scenario:

Example Scenario

The "Document Management Service" makes a chat completion API call to OpenAI. It is using the gpt-4o model, on behalf of a user, 'Jane', in order to summarize a document. This is the first step of the document summarization process, called "initial", and is one of multiple AI calls required to summarize the document as part of the "Document Summary" feature. So that the summarization can be correlated to the original document, a unique document_id is also stored. The expenses of the call accrue towards Jane's $10 monthly spend limit.

Pay-i Concept	Description	Scenario Section
Provider	The Provider is responsible for hosting an AI model and setting the price per unit. OpenAI, Azure, Anthropic, Google Vertex and AWS are example providers.	OpenAI
Category	For spend tracked in Pay-i, you can assign that spend to a Category for easy grouping. For resources whose prices are automatically managed by Pay-i (such as well known AI models), the categories are automatically defined, and begin with a "system." prefix.	system.openai
Application	The Document Management Service is an application that has multiple potential use cases.	Document Management Service
Use Case	Summarizing the document is the broad use case being achieved in the scenario.	ex. "document_summary"
Resource	A Resource is a cost generator that you want Pay-i to track. For example, `gpt-4o` is a resource because it generates a cost each time it's used. Pay-i maintains categories and resources automatically for major GenAI providers, but you can define your own.	`gpt-4o`
Request	Requests are cost-generating events that leverage one or more Resources. For example, calls to an AI model that contain a prompt and expect a completion in response.	The chat completion API call
Use Case Instance	Summarizing this specific document on behalf of Jane is an Instance of the broad use case "document_summary".	Summarizing a specific document
Use Case Step	Use Cases that take multiple actions, such as agents, can be broken down into steps so that you can track and analyze them separately.	"initial"
UserID	The user for which this request is being made.	"Jane"
Properties	Pay-i supports both custom and system managed properties. These are extra fields of information used for data pivoting and correlation.	`document_id`
Limit	Pay-i can be used for cost governance, not just observability. Limits can be used to ensure you stay within allowable margins, even for complex agentic workflows.	$10/mo

All of these terms are further explained on their own pages.