It depends on the drawing quality, the objects you count and the AI model. In our CaseV-Bench benchmark, model scores on the same architectural sheets ranged from 20% to 92%. That is why we start with a free prototype on your own drawings: you compare its counts with your own takeoff before you commit.
Get a custom prototype
We start with a short intro call. Then we take the type of drawings you work with and show you a demo of the AI counting and measuring the objects you need.
Ideal for:
- Teams looking at AI takeoff for the first time
- Testing on your own drawings before you decide
You get:
- A prototype that runs on your drawings
- Counts you can check against your own takeoff

