Bots are primarily about protocol. The reason for that is that humans are highly sensitized to protocols that are brought about by context switching. For example, observe how we change the structure of our sentences and the tone of our voice upon walking into a doctor’s office vs walking into a bank vs walking into a lawyer’s office vs walking into an airport security lineup, etc. We know that each and every situation (unless it’s highly informal, like when lounging with friends) requires a different protocol. We have no problem with that, and we’re quick to learn (survival tactics).
Conversing with bots is pretty much the same — it requires its own protocol. It would be wrongheaded to insist that, when humans converse with bots, the bots must assume the casual stance and must accommodate whatever foibles particular human may bring to the conversation.
No. The bot sets the tone in the same way that a concierge, or a customer support specialist sets the tone. That specific tone of the conversation is intended to maximize the success of the customer service.