Loading
Not ready, and the gap is documents rather than technology. Write the top twenty answers first.
Weighted so that a company with excellent technology and no maintained documents scores low, which is the honest answer.
It asks whether the preconditions for a useful assistant are in place: enough written content to answer from, a permissions model that holds, and somebody who owns the answers.
Most failed AI rollouts failed on these rather than on the model.
Answer honestly about your current state, not your intended one. How many of your top twenty questions have a written answer somewhere? When was each last reviewed? Who owns them?
The content question is usually the one that fails. An assistant with nothing to cite can only guess.
It will not tell you whether the assistant will be adopted. That depends on whether the first few answers are right, which is why the content question matters most.
It also cannot assess your data quality, only whether documented answers exist.
If fewer than half your common questions have a current written answer, write those first. Enabling an assistant before that produces refusals or invention, and both cost you.
If permissions are unclear, fix them first. An assistant applies whatever model you have, faster and more thoroughly than a person would.
Not having written answers to cite.
You need a model you understand. The assistant will apply it thoroughly.
No.
Half an hour on your own figures, and an honest answer about the parts Treepie does not improve.