How to genererate document sections based on previous documents

[levalencia]

My client has a list of 25 documents which are proposals for their clients. The documents in word have the same structure like:

1. Project Context and Summary
2. Business Activities and Project Context
3. Project Content.

Etc

They want to be able to ask the chatbot to generate a section of the document and the bot should generate with their writing style.

Example:

1. Generate a new project context and summary for the new client Cocacola and the project about improving production efficiency on plant X.

My first question is, Does this task needs to fine tune a model? If not, how would you approach this?
Do I index all document content into a vector db and then ask the questions?
Or should I preprocess and index all Project Contexts into one index, then all Business Activities sections into another index?

Thanks