AI powered document processing and automated data extraction.
Zanran Scaffolder adds structure to unstructured documents, therefore supporting and “scaffolding” them. The structure that is added – missing from a PDF or image - includes:
blocks of text – the paragraphs
headers (section headers, etc)
graphics or images
additional labelling; (e.g. for scientific papers, the following additional labels are added: abstract, citation, contributor, affiliation, and article title)
The Scaffolder can be trained to process unstructured data, from a wide range of document types, from any industry. And it does so at enterprise scale.
Annual Reports | Financial Statements | Research Reports | 10K |
10Q | Scientific Journals | Medical Notes | Clinical Trial Data |
Zanran’s approach is unique in that it uses a mix of deep learners and conventional programming to achieve optimal results.
The normal output from the Scaffolder is XML. Also available as JSON, HTML or Excel (for tables)
For a free trial - testing your reports and documents - please go to Trial
Zanran Workbench is a Windows-based tool that enables users to visualise the XML generated by the Scaffolder. As well as visualise, it provides an intuitive user interface for editing, tagging and enriching the content and its structure.
Zanran Workbench allows users to:
Quickly check, and if necessary, edit, the output from the Scaffolder.
Modify any tags that had been automatically added
Add any new tags
Change the boundaries of tables or blocks of text
Use the built-in calculator tool to verify calculations in tables
Add notes or annotations for others to see comments
For accountants: Workbench does the casting or footing checks on your draft financial reports. It picks up mistakes that people miss.
Fully utilise all of your data
Studies have shown that the quantitative data that is locked in the vast repository of unstructured documents which companies store, holds immense economic value once unlocked.
Zanran has created a proprietary AI engine to unlock the vast repository of data trapped with documents, making the data machine readable and available for search, analytics, insights and workflow automation.
Zanran’s document automation engine creates value by augmenting the RPA workflow. Their document automation engine automatically provides the required structured output data to the RPA BOTS / RPA Platforms through connectors and REST APIs for further value-add processing.
Zanran's algorithms are based on visual parameters, not linguistic or semantic ones. At present, the software works with any language that reads left-to-right and is made of letters rather than ideograms.
For a standard document of 200 – 300 pages, these complex operations are performed in just a few minutes.