Extracting unstructured document data with AI

Document AI is a product in Google Cloud Platform that that uses AI to solve industry document processing challenges.

As a hybrid product designer I collaborated with a tenured designer on the project to launch several projects in 2020.

On average, mortgage companies work with
40 to 50 document types for a loan file.

That is a lot.

But with Document AI, it's a different story.

Fully automated UX

  1. Choose processor based on document type
  2. Update document data
  3. Label sample data
  4. Auto process all documents with trained model 
  5. Review documents with low score manually
With AI models called 'processors' we allow users to automate document data redaction, validation, and processing, significantly reducing manual effort of human resources.

All users had to do was simply:

Bringing a new feature to an existing framework

I joined the project at a mid-stage where the core foundational features already existed. My role focused on establishing the Human review workflow which allowed users to set up their human review triggers, managing labeling teams and instructions. 
I also had a blast redesigning the product homepage to showcase our extended resources & features as well as the processor gallery.

Illustrations