Using Snowflake's Document AI and LLMs To Extract Data From Documents

Snowflake's Document AI (currently in Private Preview) uses a large language model (LLM) to extract data from documents. In this demo, we'll ask questions about the contents of PDF documents using natural language and get answers and confidence scores, then fine-tune the model to improve our results. We'll also see how Document AI can be used with other Snowflake features in data pipelines.

00:00 - Getting more out of your data with AI
00:21 - Demo with LLMs and Document AI
04:16 - Demo recap

Please review our legal disclaimer here:

[yjrpksq]

Can you please explain how the document has been uploaded and further process?

[talbrill]

How would one go about automating this process?

[cgovea]

Is there/will there be a feature that identifies where in the document the model found the answers?

[piotrzielonka]

That's cool!
Would you be able to tell if it is possible to process real time documents uploaded by website user and what latency to expect in such scenario?

[nulI_dev]

Is this publiclly available? I can't find any actual info other than this video

[ORIGINALwhatnow]

I gave a thumbs down because it didn't give information about how to get to this feature, how to set it up, and how the entire process works.