When adding environmental actions, the current setup is to get the exact name of things in-world, and to copy-paste into the designer. It'e technical and works for now, but it would be useful in the future to have a UI available if the bot is logged in that lists objects visible to the bot, with filters for owner and interaction type. Default owner filter would be the logged-in account that owns this bot. Default interaction filter would be Any.

Let us type an owner name, and let us choose from a drop-down of available actions for sit, touch etc.

In this way, we could easily set up the bot account to interact with the world in a more directly reliable way (using what the bot actually detects and the actions actually available).

The method in the beta is great for flexibly trying everything, but I think it will make these bots harder to set up and use for a broader audience.