Which solution to aim? (data scientists types)

I am facing this issue at some project I am working at. I'd love to open a discussion

Which solution should I aim for?

A: solution to a problem with a simple SQL query or simple plot, that will lead to an accuracy of lets say 80%

B: very complex solution (e.g. complex models and thorough representation) yealding to an accuracy of 90%

Is it worth the efford? I know, in some cases makes sense, but also in another cases, …

data datascience data scientists scientists