Ledger
Every model call is a capital outlay
Ingest raw telemetry from every provider and tie it to ERP revenue, departments and customer accounts.
Every investment you approve has to beat a minimum return, except AI. Hurdle r8 measures every model, team and customer against your hurdle rate, straight from token telemetry to the general ledger.
AI Portfolio · Q4 2026
Return vs. 12% hurdle rate
| Initiative | Spend $ | Return |
|---|---|---|
Customer product — Search Anthropic | 184,210 | ▲ 31.4% |
Support agents OpenAI | 41,905 | ▲ 18.2% |
Internal workflows Azure OpenAI | 27,330 | ▼ 9.6% |
Coding assistants Cursor / Copilot | 62,480 | ▼ 4.1% |
Sample data. ▲ clears hurdle · ▼ below hurdle.
“6 out of 7 CFOs don't know the ROI on their AI spend, and the 7th is lying.”
Why Hurdle r8
Ledger
Ingest raw telemetry from every provider and tie it to ERP revenue, departments and customer accounts.
Returns
Separate customer features from coding assistants and internal tools. See what earns its cost of capital.
Governance
Budget owners hear when an initiative slips under the hurdle, not at month-end close.
Audit
Amortize prepaid credit pools against burndown. Telemetry ties to AP invoices so auditors sign off.
How it works
Use your WACC, a target IRR, or a rate per business unit.
OpenAI, Anthropic, AWS Bedrock, Azure OpenAI and AI gateways. No engineering sprint.
Fund what clears, cut what doesn't. Ask Claude or Copilot through MCP.
“We stopped debating AI budgets as a cost line. Now every initiative has to show its return against the hurdle, same as any other capex.”
“The only platform that reconciles credit burndowns to the individual token and ties them back to our GL for auditors.”
“It bridges engineering logs and FP&A reporting without forcing developers into manual tagging spreadsheets.”
Connect one provider. Within 24 hours you get a reconciled AI portfolio, every initiative measured against your hurdle rate.