Overview
Use AI Usage Data to load, review, and manage monthly AI consumption data for cost analysis and reporting in Yarken.
You can upload AI usage files manually for individual months or load supported data automatically through Advanced Integrations. Once loaded, the data is available for reporting in Analytics > AI Consumption.
The page provides a monthly overview of uploaded AI usage, including spend and the most recent upload date.
Note:
AI usage data must be loaded, either manually or through a supported Advanced Integration, before it can be analyzed in the AI Consumption cube under Analytics.
Standardized Cursor AI usage data
Cursor usage data imported through Advanced Integration uses the following standardized field values for consistent AI cost reporting:
|
Field |
Value |
|---|---|
|
Service category |
AI and Machine Learning |
|
Publisher and provider |
Cursor |
|
Service name |
Cursor |
|
Collector source |
cursor-admin-api |
|
Charge category |
Usage |
|
Billing account ID |
<yourorganizationname>-Cursor |
These values help ensure that Cursor usage is categorized consistently and is available for cost analysis in the AI Consumption Analytics cube.
Before you upload AI usage data
Ensure that:
-
Your AI usage file matches the supported data format.
-
The reporting year is selected before uploading data.
If your organization uses a supported Advanced Integration, Yarken can retrieve and load AI usage records automatically. The file format and manual upload requirements above apply when you upload data yourself.
AI Usage overview
The page displays one card for each month of the selected year.
Each monthly card provides:
|
Item |
Description |
|---|---|
|
Month |
The reporting month. |
|
Spend |
Total AI spend uploaded for the month. |
|
Usage indicator |
Visual representation showing that AI usage data has been loaded. |
|
Last Load |
Date and time of the most recent successful upload. |
|
More options |
Access additional actions for the selected month. |
Months without uploaded data display a spend value of $0 and no last load information.
Upload AI usage data
To manually upload AI usage data:
-
Navigate to Designer > Consumption > AI.
-
Select the required reporting year.
-
Locate the month you want to upload data for.
-
Open the More options menu on the monthly card.
-
Select Upload.
-
Choose the AI usage data file.
-
Select an existing Data Mapping Template, where applicable.
-
Map the source fields in your file to the corresponding AI Usage Data fields.
-
Review the mappings and continue through the upload process.
-
Complete the upload process.
After the upload completes successfully:
-
The monthly spend is updated.
-
The Last Load timestamp is refreshed.
-
The uploaded data becomes available for reporting in the AI Consumption Analytics cube.
TBM and allocation dimensions
AI Usage records can also include TBM and business allocation dimensions.
These fields can be populated during AI Usage Data ingestion, including through supported Upload Rules, and are available for reporting where populated.
|
Attribute |
Description |
|---|---|
|
Cost Pool |
Cost Pool classification associated with the AI consumption. |
|
Sub Cost Pool |
Sub Cost Pool classification associated with the AI consumption. |
|
Tower |
Tower classification associated with the AI consumption. |
|
Sub Tower |
Sub Tower classification associated with the AI consumption. |
|
Solution |
Solution associated with the AI consumption. |
|
Consumer |
Consumer associated with the AI consumption. |
Automated ingestion through Advanced Integrations
If a supported Advanced Integration is configured for your environment, Yarken can automatically collect and load the provider's AI usage data. You do not need to manually upload the same records.
Yarken manages the integration and transformation workflow. Your organization authorizes the provider connection and uses the Advanced pipeline to run or monitor data collection. The processed records are loaded into AI Usage Data under Designer > Consumption > AI.
Automatic ingestion provides:
-
Scheduled data ingestion.
-
Reduced manual effort.
-
Consistent monthly AI usage reporting.
-
Faster availability of data for Analytics.
For setup, connection authorization, and pipeline management, see Advanced integrations. For provider-specific prerequisites and data behavior, refer to the guides in the table below.
Supported Advanced Integrations for AI Usage Data
The following supported Advanced Integrations load usage and cost records into AI Usage Data:
|
Advanced Integration |
Supported data loaded into AI Usage Data |
|---|---|
|
Claude Enterprise user-attributed usage, tokens, costs, and supported model, product, and team attributes. |
|
|
Anthropic API costs, token consumption, workspace usage, user and API key attribution, and supported tool charges. |
|
|
OpenAI API Platform organization usage, cost, and tokens, including project or user attribution where available. This integration does not collect ChatGPT Workspace usage. |
|
|
Daily ChatGPT Business and Enterprise workspace activity by user and client (Chat, Codex, and Agents), including credits, input, cached input, output and total tokens, main model, and usage-based costs. Seat subscription fees are not included. |
|
|
Daily Copilot usage by user, model, and feature, including AI credits, estimated cost, and supported token metrics. |
|
|
Supported user and model usage, token consumption, and associated costs. |
|
|
AI request usage, tokens, and costs, with provider, model, and supported user, team, and tag metadata. |
Data format and availability: Advanced Integrations can retrieve source records through APIs instead of transferring physical upload files. Available fields, reporting periods, and data freshness depend on the provider and integration. Not every source provides the same file format or set of metrics.
Review automatically ingested AI usage data
After an Advanced Integration is configured, validate the imported data:
-
Go to Admin > Pipelines > Connections, select the ADVANCED tab, and confirm the provider connection shows Connected.
-
Go to Admin > Pipelines > Pipelines, select the ADVANCED tab, and check the pipeline status and run history.
-
Go to Designer > Consumption > AI and select the relevant reporting year and month.
-
Review Spend and Last Load to confirm data was loaded for the expected period.
-
Go to Analytics > AI Consumption to review imported provider, user, model, cost, and usage data where supported by the source.
-
If data is missing or does not match the source, review the pipeline history and the provider-specific guide for availability, limitations, and troubleshooting guidance.
Upload rules
Select the Upload Rules tab to configure or manage the rules used when importing AI usage data.
Populate fields with Upload Rules
You can also use Upload Rules to populate supported AI Consumption fields during ingestion. Use Upload Rules when the required classification value is not provided directly in the source file or when you want to apply a consistent value based on rule conditions.
Publish AI cost and usage data
After AI cost and usage data is loaded, you can publish the monthly spend to one or more Spend models.
Publishing creates corresponding records in Expenses for the selected models.
Note: AI usage publishing currently stops at Expenses. Yarken does not distribute the published spend through Cost Pools or Towers, and publishing does not create Assets.
Publish AI spend
-
Go to Designer > Consumption > AI.
-
Select the required reporting year and month.
-
Open the AI usage data for the month.
-
Select PUBLISH.
-
In AI Summary, select the Spend models that should receive the AI spend.
Select the Spend models that should receive the AI spend. -
Review your selections.
-
Select PUBLISH SPEND.
Yarken creates the corresponding records under Expenses in each selected Spend model.
What happens after publishing
Published AI usage records:
-
are created at the Expenses level,
-
are not distributed to Cost Pools or Towers,
-
do not create Assets, and
-
retain their AI origin in the published Expense record.
Published AI usage records use the following source classification:
|
Field |
Value |
|---|---|
|
Source |
AI |
When you publish AI spend from Designer > Consumption > AI, Yarken creates the corresponding Expense records in the selected Spend models with Source = AI.
You can use the Source field to identify AI-related Expense records and distinguish them from Cloud-published spend in reporting and analysis.
Note: The Source field is non-editable. Yarken automatically assigns AI when publishing spend from AI Usage Data.
Current publishing scope
|
Capability |
AI usage publishing |
|---|---|
|
Select Spend model |
Supported |
|
Publish to Expenses |
Supported |
|
Cost Pool distribution |
Not supported |
|
Tower distribution |
Not supported |
|
Asset creation |
Not supported |
Unpublish AI spend
You can unpublish AI cost and usage data using a workflow similar to Cloud publishing.
-
Open the published AI usage data for the required month.
-
Select the option to unpublish the AI spend.
-
Confirm the action.
Yarken removes the previously published AI usage records from the applicable Spend models.
Best practices
-
Keep AI usage data current by uploading files when needed or using supported Advanced Integrations for recurring ingestion.
-
Verify the reporting month before uploading a file.
-
Review Last Load to confirm the latest successful data load.
-
Use Advanced Integration for recurring AI usage imports to minimize manual administration.
-
Validate loaded data in AI Consumption after manual uploads or Advanced pipeline runs.
Related content