Looking for AI model recommendations for generating documents based on a library of previous documents

[axipher]

I'm looking to dip my feet in some more useful Generative AI than just using ChatGPT for basic Code generation.

 

My work is smartly taking a cautious approach to implementing AI as tools for employees to use.

 

One part of my work is to generate Proposal Documents for potential jobs (new conveyor belt for an underground copper mine, new rock crusher installation, new air quality monitoring and automated ventilation, etc.)  We have thousands of existing Proposals that just get recycled but always end up being pretty narrow-minded.  So one of our AI initiatives was to look for something that can run either local, or in our already paid for Azure Cloud, that we can feed in our massive inventory of existing proposal documents.

 

Ideally:

- Use local storage, or known secure cloud storage with known datacenter selection for legal requirements

- Extract basic meta data from title page of proposal (word documents)

- Be able to rank proposals based on if they were successful or not

- Take some prompt questions about a new job

- Search through existing proposals for seemingly related work

- Generate a new Proposal that can be used as a baseline for creating something new

 

 

I feel like there are so many options out there now, and I just don't have the AI know-how to try to navigate the different types of models, ways to convert our existing documents in to tokens for the AI model, etc.  I'm really hoping to leverage some of the knowledge here to guide me in the right direction for some options to look at testing out.

[UltraMega]

1 hour ago, axipher said:

I'm looking to dip my feet in some more useful Generative AI than just using ChatGPT for basic Code generation.

 

My work is smartly taking a cautious approach to implementing AI as tools for employees to use.

 

One part of my work is to generate Proposal Documents for potential jobs (new conveyor belt for an underground copper mine, new rock crusher installation, new air quality monitoring and automated ventilation, etc.)  We have thousands of existing Proposals that just get recycled but always end up being pretty narrow-minded.  So one of our AI initiatives was to look for something that can run either local, or in our already paid for Azure Cloud, that we can feed in our massive inventory of existing proposal documents.

 

Ideally:

- Use local storage, or known secure cloud storage with known datacenter selection for legal requirements

- Extract basic meta data from title page of proposal (word documents)

- Be able to rank proposals based on if they were successful or not

- Take some prompt questions about a new job

- Search through existing proposals for seemingly related work

- Generate a new Proposal that can be used as a baseline for creating something new

 

 

I feel like there are so many options out there now, and I just don't have the AI know-how to try to navigate the different types of models, ways to convert our existing documents in to tokens for the AI model, etc.  I'm really hoping to leverage some of the knowledge here to guide me in the right direction for some options to look at testing out.

 

 

I think this would work https://blogs.nvidia.com/blog/chat-with-rtx-available-now/

 

 

Closest thing I'm aware of to what your looking for, tho there are probably dozen of other options.