Overview
Duco’s adaptive intelligent document processing (IDP) technology extracts unstructured data from documents, emails, images and other unstructured formats. But unlike traditional IDP solutions, it doesn’t follow a rigid template to recognise document or data types. It uses a deep learning engine with proprietary NLP algorithms and AI models that continuously improve with human validation and enrichment. Business users ‘train’ the machine simply by doing their job - no AI skills necessary.
How it works:
- Pre-processing: optimises data and prepares for extraction
- Classification: Splits files, recognises and classifies document types, and paginates as appropriate
- Extraction: extracts data formatted to your configuration; links connected text (eg payment amounts and dates) and recognises signatures, stamps, logos
- Validation: validates data against business rules and custom built enrichment with external data
- Output: sends to data lakes, further Duco processes, or other applications
Key Features
Humans in the loop: manage how and when humans feed back to train and maintain deep learning models
No code: business users don’t need code or AI skills to set up or to train/maintain the model, removing dependencies on overstretched IT resources.
Configurable and customisable: add, modify and enrich in platform; adapt process pipeline with ready-built, customisable code snippets
Detailed production and model analytics: provide tracking, insights and potential improvements
Key Benefits
Handles unlimited variety - emails, documents, scans - and does not need to be retrained when a format changes
Can be used across the business, eliminating point solutions and consolidating the tech stack
Self-learning models improve over time, increasing STP rates and efficiency
Reduces human error and associated risks