What is Document AI?

Welcome to The Future of Documents, a new series that helps you make the most of your unstructured data with Document AI. In this episode, Developer Advocate Holt Skinner shares how there’s an enormous amount of data in the documents we interact with, such as PDFs and emails, that contain unstructured or “dark” data. Join us for this exciting new series where we show how you can use this data to automate processes, gather analytics, and much more!

Chapters:
0:00 - Intro: A goldmine of data
0:57 - Document understanding with AI/ML
2:18 - What is document AI?
3:08 - Document AI building blocks
3:55 - Recap / Wrap up

​​
#TheFutureOfDocuments

[jaisonleon]

Tried document AI to parse 77 page pdf files. Works great. Would be great to make it compatible with big query by removing empty structures.

[jmxt]

The instructor is so colorful, like a rainbow

[nikolai]

IBM did this 5 years+ ago with "Watson" but it is cool that it is made accessible for everyone

[KManAbout]

Can't wait to use this. Going to try this on some of the survey data I got.

[markfortner]

This looks really interesting. It would be nice to be able to parse journal articles and extract gene/protein information, pathway information and drug information. EuropePMC has a large set of free text papers that could be used to train a model for this.

[develiorate]

Thank you for the interesting 🤔 infos
Google doing well 😊👍

[TheBoxMover]

Hello,
Can "Document AI" also be used to process texts from PID drawings in A0 format where the texts are horizontal, vertical or at any angle?

[ivanferreira]

Really cool application, does it work for languages other than english (e.g. spanish)?

[Josephkerr]

I have a bunch of journal entries and sketches is there going to be a way to create a digital twin of my sketch pages and pair journal notes to each? Then I could hyperlink pages pertaining to similar content and organize notes better. They are all old notes, and while I've transitioned to digital documents for writing, there is still issues with bringing scans to those documents. Very time consuming to structure it manually.

[claygli]

Can it work with handwritten documents? Does it understand handwriting?

[arjunabeuger]

Dark Data sounds cooler than Document AI

[jessejames]

awesome!!!! so eexcited to learn more thank you Google!!!

[LucidSonder]

I thought this was a model that documents code for you. Would be a fun challenge.

[orly]

FINALLY, this is huge, you guys doing good work

[Kudz_LastmanStanding]

This is absolutely sick, 🔥🔥
This will make me rich lol, the amount of apps i can build of this feature is crazy!
GCP is ahead of the game with AI

[sitrakaforler]

Dam cool! I find it so...crazy to have everything on the cloud but it's indeed usefull for some works :D

[ErcoleBellucci]

This is AI talking presenting us Document AI

[janberger]

The volume is insanely high compared to YT average

[fslurrehman]

Is doc ai same as knowledge graph + ml . I am confused here. Can Google advocates answer this?

[st]

Very interesting and useful technology. But I can see ways in which certain... authoritarian governments... could benefit very much so from this.