Maintain user agency as the AI works in order to understand and direct the AI's logic
Lets the AI give inline annotations to cite its sources
Manage the flow of information or pause a request mid-stream to adjust the prompt
Let users trace the AI's steps from prompt to result
Show users what is actually happening behind the scenes
Have the AI reproduce its response to the prompt without additional input
Confirm the user's intent for complicated prompts
Have the AI show the steps it will take to respond to the user's prompt before it executes its response
Reveal the tokens the AI used to craft its response
Trace through multiple variations of a result to choose the one that works best for them