Why not just have a touchscreen menu then? You already need large screens so people can confirm the AI recorded their order correctly and this will skip the need of a person manning the drive through menu. You could even include options to "hold the pickle", etc.
This would also be nice. Usually, I only order fast food if I can place a call first (Indian, Chinese, Mexican, Pizza, etc). Grubhub fees are ridiculous or and apps are always taking your data.
The issue with touchscreen kiosks is that some have short arms.
On another note, I get the benefits of computer-voice-operated drive-thrus. No need to use your phone. If your phone's 🔋 is 3%, you can still buy food.