Automate the boring work

Don’t waste valuable human capital on manual document processing. Duco’s adaptive IDP can help you reduce operational costs and cut the time you spend on manual data extraction.

Model training, evaluation, version management, deployment, configuration, input/output, user management, are all built-in.

No need for coding, no AI skills required.


Adaptive IDP:

Adapts and learns through integrated human feedback

Automates data extraction on any document type – including custom formats – and any kind of data.

Enables you to build models with as few as 10 examples, or adapt an existing model from our library

Enables process customisation and in-platform data enrichment

Adaptive IDP for unstructured data

Adaptive intelligent document processing (IDP) uses a deep learning engine: proprietary NLP algorithms power AI models, continuously improving with human validation, enrichment and threshold scores


Optimise and prepare for extraction


Split files, recognise/classify doc type, paginate


Extract data formatted to your configurations. Links connected text (eg dates and amounts), and recognises signatures, stamps, logos…


Validate against business rules and enrich with custom logic


Send to further Duco processes or other applications

The adaptive difference

Integrated features to help you extract unstructured data with ease.

Fast, easy labelling

Clear, straightforward guidelines for labelling guidelines and step-by-step instructions to help business users label document elements and help train the model, and improve it in production.

In the training phase, the model identifies and suggests the best documents to label, reducing effort while improving accuracy. It will also suggest where it thinks business users have made mistakes in labelling, speeding review.

Machine learing

AI + humans

Machine learing

Adaptive IDP enables full configuration of where and when humans provide feedback to help improve the deep learning models.

Feedback from business users as they do their daily work helps train the models without the need for specialist skills. Models learn from interactions and improve results.

Fully customisable

  • Easily add document types, document elements, enrichment or validation rules
  • Validate and enrich with business rules
  • Customise how the extraction process works based on your business requirements

Queueing and monitoring

Real time view of AIDP processes enables business users to see what’s happening with the model and identify opportunities to collaborate with other team members.

Real time dashboards monitor performance to keep you on top of cost and STP rates. You can also see STP rates, data quality and other model analytics to help identify potential improvements to the model.

Analytics enable users to track, compare and optimise the model from training to production.

Manual & legacy solutions

Manual processing is slow, prone to errors and costly

OCR template-based solutions require structured input with fixed layouts

Developers or consultants manually create, maintain and adjust templates whenever formats change – costly and time consuming

Point solutions for each document type/department create a messy, expensive tech stack

Pre-trained models deliver mediocre, static performance

WITH adaptive IDP

AI handles unstructured input with unlimited variety (documents, scans, emails, etc.)

Algorithms learn from context and interpret layout and text, not just memorize it

Transparent training models – no black box

One solution used across the business – no specialised point solutions

Self-learning models provide better performance and improve over time

Adaptive IDP is part of the Duco platform

Agile & scalable

Harness the power of the cloud to react fast to any changes in data volume or complexity.


Easily automate common tasks with high accuracy, keeping humans in the loop to validate and improve outcomes


Standardise and control your data across your organisation.