AI spend tracking dashboard for Cursor, Claude, OpenAI, and AWS Bedrock

By Nico Waissbluth

Works with

Try this example in Tines 3B

This workflow consolidates AI spending data from multiple platforms (Cursor, Claude, OpenAI, AWS Bedrock) into a unified dashboard with per-user and per-department attribution. It fetches billing data hourly from provider APIs, normalizes it alongside HRIS employee data, and serves an interactive React dashboard with time-series trends, breakdowns by user/model/source, and exploratory visualizations. Finance and IT teams can track aggregate AI costs, identify usage patterns, and analyze model adoption across the organization.

The problem

AI spending is spread across provider dashboards, billing reports, and shared accounts, making it difficult for Finance and IT teams to understand where costs are growing. Different reporting formats and incomplete employee attribution make comparisons harder, while manually combining exports takes time and can leave gaps in the picture.

Without a shared view of AI usage and costs, teams struggle to identify spending trends, plan budgets, and follow up with the right person when usage exceeds expectations. Billing delays and estimated costs can also create false confidence if they are presented as complete, final charges.

Solution and impact

This workflow brings usage and spend data from Cursor, Claude through Anthropic’s API, OpenAI, and AWS Bedrock into a shared AI spend tracking dashboard. Scheduled collection and automated reporting replace recurring spreadsheet consolidation, while an optional employee directory integration connects usage to names and departments.

Finance and IT teams can filter spending by date, provider, employee, department, and model, compare usage patterns, and export filtered data for further analysis. Optional forecasts support budget planning, while visible coverage and freshness warnings distinguish missing data from zero usage. Reported, allocated, and estimated costs remain clearly identified rather than being presented as exact employee-level bills.

Optional Slack alerts flag daily spending that reaches a configurable threshold. Employees receive a link to review and explicitly acknowledge their spending, with reminders and manager escalation for unacknowledged alerts. A separate alert history view records delivery and acknowledgement activity.

The result is a clearer view of AI costs, less manual reporting, and more consistent follow-up on high spending. Teams can make better-informed budget decisions and identify opportunities to manage costs without automatically restricting access to AI tools.