Short answer - is it too early to give any solid answer.
It is designed to be fully LLM editable with full control over changes. So I target the use case when any change requested calculate whole chain of the impact for the review.
Right now it can rewrite the document in a specific way on follow-up promts.
This rewriting uses gpt4o tool calls (so kind of basic reasoning), over the weeked will put Gemini Pro 2.5 to that editing service (then going to try grok 3). My expectation is that its going to feel like your doc got edited by chief architect (gemini pro) vs junior pm (gpt4o)
It is designed to be fully LLM editable with full control over changes. So I target the use case when any change requested calculate whole chain of the impact for the review.
Right now it can rewrite the document in a specific way on follow-up promts. This rewriting uses gpt4o tool calls (so kind of basic reasoning), over the weeked will put Gemini Pro 2.5 to that editing service (then going to try grok 3). My expectation is that its going to feel like your doc got edited by chief architect (gemini pro) vs junior pm (gpt4o)