Does it actually work?
We tested 9 frontier AI models on 43 real conveyancing scenarios — first with no help (baseline), then equipped with our open-source skills. The results speak for themselves.
Overall Accuracy: Baseline vs With Skills
43 test cases across SDLT, Lease Impact, Lender Pre-Screen, Restrictive Covenants, and Title Defect Analysis
| Model | Baseline | With Skills | Improvement |
|---|
Breakdown by Skill
Format: Baseline% → With Skill%
| Model | SDLT Calculator | Lease Impact | Lender Pre-Screen | Covenants | Title Defect |
|---|
Last updated: · View methodology & raw data