The main difference is in the linguistic complexity.
People express themselves differently when they speak compared to when they type. When we speak we use more sentences and we make our sentences longer.
As a result a voice bot needs to have better AI compared to a chatbot, in order to handle a conversation and deliver the same customer experience.
If your business model allows it, is better to start with a chatbot and add a voice bot on top of it.
This way you can gradually increase the complexity of your AI without compromising on your customer experience.