How Johnnie built a custom ARR engine in just a few hours

Johnnie Walker, Director at Rooled, wanted ARR reporting calculated his way – the way finance leaders actually think about revenue.

Using Rillet’s newly released Model Context Protocol (MCP) and Anthropic’s Claude Code, Johnnie built a custom ARR reporting engine in just a few hours. 

What he built

A local, self-contained application that:

• Pulls live customer, contract, and invoice data directly from Rillet
• Applies clearly defined ARR reporting policies
• Produces detailed per-period and per-customer exports
• Identifies inconsistencies automatically
• Reconciles reported ARR with recognized revenue
• Exports cleanly to Excel for validation and analysis

If something didn’t look right, Johnnie could iterate on the logic and lock the fix in with a test, treating finance logic like software.

How Rillet MCP made this possible

Model Context Protocol (MCP) allows AI tools to securely interact with external systems.

With Rillet MCP, Claude Code could:

• Query Rillet’s APIs directly
• Search Rillet documentation as needed
• Generate code grounded in Rooled’s real financial data

Furthermore, this lets finance leaders focus on judgment and policy, while the AI handles the underlying data work and implementation.

The Takeaway

As AI reduces the cost of building, the real advantage comes from domain expertise. Rillet MCP gives finance leaders like Johnnie the ability to turn that expertise into working systems. Fast, transparent, and fully under their control.
‍

Read his experience here.

‍