By clicking "Accept all", you agree to the storing of cookies to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our privacy policy for more information.

Technology

Futureproof and scalable technology

Read More Icon
read more

Conpend’s AI apps use revolutionary data capture technology, pairing big data concepts with AI to extract data from unstructured text. A real-time feedback loop enables progressive machine learning, resulting in exceptional accuracy without prior knowledge of the documents.

All Conpend’s AI apps are designed to be both future proof, with no legacy components or workarounds, and highly scalable. They also have client security in mind, from design to operations. Conpend’s platform is based on Microservice and API architecture and can be hosted with any leading cloud provider, as well as on premises or in a colocation centre.

Our Features

OCR (Optical Character Recognition)
OCR based on ABBYY FineReader. ABBYY is commercially part of the solution and as such Conpend will provide support on ABBYY products.
Standard license is for Active / Passive Production configuration. Active/Active license is available on request.
OCR Vendor independent
The Trade AI system can be configured to work with any OCR service (such as Captiva, Kofax or FileNet) as long as such a service provides output with sufficient detail (assessment on request).
AI driven False Positive suppression
Automatic False Positive suppression enabled in Trade AI
Document classification
Automatic classification of all documents
Reading detection
Copy / Original
Data Extraction 100% AI driven
Automatic learning data extraction
ICC checks
400+ Business checks
Document preparation
Covered by EssDOCS
TBML checks
200+ Business checks
Vessel checks
Lloyds list standard interface
Price checks
World economic forum standard interface
Straight Through Processing
Risk based
Microservices with open API’s
Adapting to your infrastructure with Microservices architecture and Open API methode
Advanced customizable workflows
4 eyes principle, conditional escalations
Native integration Finastra Fusion
Automatic interface enabled with Finastra (FBTI patch 2.33 and higher)

Digitization: The Automation Solution

Document Recognition and Data Capture - Escaping Templates

Trade AI uses revolutionary data capture technology which departs from the traditional template approach and uses modern big data concepts paired with Artificial Intelligence (AI) to accurately extract data from unstructured text. Using an evolving data set and user supervision a real-time feedback loop is created, similar to bias units in a neural network, which enables progressive machine learning resulting in exceptional accuracy without prior knowledge of the documents.

Trade finance challenge
Our Trade solution
Low repetition in layout
Data extraction is data driven based on a multi-dimensional machine learning model which learns from user interaction and gets smarter the more it is used
Required values not always predictable
Automatic identification and extraction of relevant values based on the Letter of Credit which is dynamically adapted based on the transaction contract and can be refined by the user
Document Set structure is critical
Document sets are consolidated into the key “master” documents using In Document set Learning (IDL) reducing the time spent on extraction and validation
Close relationship between data extraction and checking / screening
Data extraction and transaction screening integrated into one system allowing simple context specific checks which can also be initiated through an AP
Transaction processing is under constant time pressure
Users can handle all exceptions on the fly within the system without needing intervention from an engineer. All exception handling is automatically added to machine learning dataset