Leon Richter
Support Specialist · Customer Support
LR
ROI score
69Watch
Premium model used where a cheaper one would do.
Recommendation
Set a cheaper default model and escalate to premium explicitly.
AI spend
€80
10% vs prior period
Yield rate
56%
Team average 61%
Cost per accepted unit
€1.63
Team average €3.44
Leverage vs last month
0.91
88 runs · 49 accepted · 1.6M tokens
Weekly AI spend
Human loops
Accepted81%
Rejected 19% · team 73% accepted
Guidance levelLow
Runs without result31%
Waste€64
ROI score breakdown
Yield 30%
56
Cost efficiency 25%
100
Trend 20%
43
Model fit 10%
36
Capability use 15%
100
Each component is scored 0–100; 50 means on par with the team median or flat vs last month. Good ≥ 70, Watch 40–69, Bad < 40.
Model mix: spend share vs token share
Premium reasoning model78% spend · 27% tokens
Standard model18% spend · 13% tokens
Fast small model5% spend · 60% tokens
Top skills used
- Drafting
- Spreadsheet automation
- Image generation
Agents and MCP usage
- Spreadsheet agent149 runs
- Knowledge base MCP176 runs