Help with building a single long form document with one workflow and multiple nodes providing input
I want to build a client or lead analysis document that provides an automated analysis on a business or prospect providing valuable data in long form analyst doc. This doc is pieced together by different specialist functions or api calls to LLM's that create parts of the document on step at a time which response data is saved somehow for the workflow to retrieve and use as it proceeds along, eventually the final document is pieced together from the different parts created and a final analysis and tweaks is done by a top tier LLM quality assurance agent/assistant.
The point is through a multi step workflow run can i gather data and output from different sources during the workflow which is then pieced together at the end (prob via all parts being sent to a long content llm) to create a long form analysis document that is the final output posted as a word document somewhere or saved to google drive or emailed that was based on and created by multiple actions that occurred before and previously saved to DB data or cached by the workflow.
Essentially, I just want to know if this is easy and possible to do with buildship and the workflow has a database or storage capacity or large cache, vector and normal ones to save data to (or is that not build in DB is another api call?) and that it can work with large chunks of text being passed around, any roadblocks or problems with this idea, i need to know first before i spend time learning and testing as im on the clock, just need to know if its possible and "easily done" with no potential issues. Thank you so much.
The point is through a multi step workflow run can i gather data and output from different sources during the workflow which is then pieced together at the end (prob via all parts being sent to a long content llm) to create a long form analysis document that is the final output posted as a word document somewhere or saved to google drive or emailed that was based on and created by multiple actions that occurred before and previously saved to DB data or cached by the workflow.
Essentially, I just want to know if this is easy and possible to do with buildship and the workflow has a database or storage capacity or large cache, vector and normal ones to save data to (or is that not build in DB is another api call?) and that it can work with large chunks of text being passed around, any roadblocks or problems with this idea, i need to know first before i spend time learning and testing as im on the clock, just need to know if its possible and "easily done" with no potential issues. Thank you so much.
