Latency 3 times longer than the estimate

Example of call: 03b7b32e-ca04-47a2-ba3f-e6687cd87ae1

I've been using Vapi for a few months now and I've never managed to get anywhere near the estimate (lots of different combinations with the different providers).
There's a huge difference between the estimate and the true latency.
This latency has a huge impact on the user experience and the validation of a product.
How do you explain this?

Also, could coding the solution and hosting it on a fast cloud server in France reduce latency?

Thanks!
Was this page helpful?