The customer wanted to build an auditing portal that used machine learning to reduce the time taken to review large documents. The solution required building a user interface that allowed the auditors to navigate through classification fields, review the excerpts in the document related to each classification field and interact with the document as they would using Adobe PDF viewer.
Auditors should be able to upload upwards of 500 documents at a time while still auditing other documents.
Auditor should be able to perform actions such as select, copy and paste text from documents that are uploaded. The text layer had noise and needed to be superimposed on each document.
Memory footprint and compute needed to be conserved such that the product would work on a machine with low configurations. Lazy loading and recycling needed to be implemented for document viewing.
Document would have some pages in landscape and some in portrait. For a lazy loaded document scrolling to a specific page or excerpt required computing the page size dynamically.
Interactions similar to Adobe PDF viewer were required which included features like quick searching of lazy loaded documents, zoom in/out, fit-to-screen and scroll to page.
Wednesday worked with a partner company in a staff augmentation capacity.
The team built a solution that allowed the auditor to link master contracts to amendments with a custom and user friendly drag and drop interaction.
The team built plug and play UI components for the PDF viewer, document upload and virtual scroll.
The product was capable of being white-labelled.
Auditors saw a decrease in average time needed to review a document.
Auditors were able to link master documents to amendments easily.
Auditors were able to add new classification fields.
Talk to us
Psst! Listen to our podcast The Wednesday Show here ↗