# Global Private Banking

How a global private bank calibrated AI tool investment with per-team impact metrics.

---

## The Challenge

The bank had paid for Copilot seats for a year without knowing if they were worth it. Engineering leadership needed to justify AI tooling spend to finance and risk, and had no per-team or per-tool visibility into actual impact.

## The Solution

iftrue deployed on-prem inside the bank's private cloud. AI code ratio, churn, and prompt-to-merge are now tracked per team and per AI assistant. Leadership sponsors the tools that prove their worth, and sunsets the ones that do not.

## The Results

First full quarter after on-prem rollout.

### AI code ratio

+183%  
Before 12%  
After iftrue 34%  
Measured AI contribution to merged code, once attribution was on.

### Code churn

-50%  
Before 14%  
After iftrue 7%  
Share of merged code re-written within 3 weeks. Halved with better AI review flows.

### Prompt to merge

-41%  
Before 5.4h  
After iftrue 3.2h  
Median time from first AI prompt to merged PR across all tracked repos.

---

## See what your engineering org is missing.

30-minute walkthrough, tailored to your stack. No slides. Your team's data.