Product / Coming soon
From the tool
to the work.
AI spending is scattered. The work it supports is scattered too. ai-cur brings the two into the same frame.
Start with usage and cost. See where capacity goes unused. Then connect the spend to the people, teams and outputs behind it.
Cost & visibility
One view.
Different kinds of cost.
A paid subscription, an API charge and a usage estimate are not the same thing. Your report should not pretend they are.
Bring provider bills and workstation usage together. Follow a charge by provider, model, person or team. Keep settled billing separate from live estimates, and leave missing prices visible.
- What was paid
- Billing evidence, periods and currencies.
- What was used
- Observed usage, quota or credit information where the source supplies it.
- What is missing
- Unavailable prices, incomplete coverage and stale sources, not invented zeros.
Usage & opportunity
Before buying more,
understand what gets used.
Personal usage answers “what do I have left?” Team reporting asks “what are we paying for?”
The local collector keeps workstation visibility close to the person using the tools. The business layer brings together declared cost sources and team mapping so a lead can examine seat activity, plan fit and budget variance.
- Unused capacity
- Compare paid seats with measured activity. Review the evidence before removing a seat.
- Plan fit
- Compare subscription cost and usage only when there is a valid price basis.
- Budget position
- Follow spend by team and period. Reporting is distinct from gateway-enforced limits.
Work & return
The invoice tells you the cost.
The work gives it context.
A shared ontology connects cost and usage to people, teams, projects and the work they deliver.
Work systems record different things: a merged pull request, a resolved issue, a closed case, a commercial outcome. Keep those meanings distinct. Join them through declared identities, periods and ownership, then apply deterministic allocation rules.
- Resolve the context.
Map cost and work records to people, teams and projects. Keep unmatched records visible.
- Declare the allocation.
Show how spend was assigned, which period it covers and which cost basis was used.
- Measure comparable work.
Read cost per output and throughput with the denominator, coverage and link strength beside them.
Cost per output supports an ROI decision. It does not, by itself, prove that AI caused an improvement or establish financial return.
Personal to enterprise
More context,
not more noise.
- Personal
- Usage, quotas and limits across your tools. GitHub and Trello are examples of work context.
- Team & Business
- Billing sources, team mapping, budgets and shared delivery. Jira, GitHub and Slack add different kinds of context.
- Enterprise
- Organization, service and commercial records, including ServiceNow, Salesforce and enterprise data held in Databricks or other systems.
These examples include available and planned integrations. Slack activity is context, not a count of completed work. Check connector coverage.
Collection & reporting
Local collection.
A separate business layer.
The workstation collector is not the company database. It provides consent-aware local visibility. The business layer ingests declared sources, resolves organization mapping, allocates cost and serves reports within the viewer's access scope.
The portfolio includes standalone, Databricks and ServiceNow profiles. Delivery and connector availability differ by profile; they are not interchangeable runtimes.
Coming soon
Tell us what
you need to see.
Your tools, your cost sources, your team structure, and the question behind the report.
Contact for more details