[Docs] Improvements to the information architecture

[stichbury]

Description

I had this as a ticket and am creating it over here in Kedro as it makes more sense. Have discussed with @rashidakanchwala

Idea is to:

• Create one or more draft information architecture plans for Kedro docs that follow basic diataxis principles of different user goals and knowledge acquisition or task activation.
• Then use the draft IA, Kedro's llms-full.txt and gen AI to attempt to rebuild Kedro docs along the new IA.
• Experiment with the different IAs and try different LLMs to see which builds docs that are most convincing at flattening the curve.
• Consider how to get feedback on it, e.g. should we create a subproject of Kedro that contains the new docs and publish it?

Documentation page (if applicable)

Context