Skip to main content

SaaS platform for automated categorisation and mapping of digitised documents using machine intelligence semantic extraction

Periodic Reporting for period 2 - OMNIUS (SaaS platform for automated categorisation and mapping of digitised documents using machine intelligence semantic extraction)

Reporting period: 2020-03-01 to 2021-02-28

The Problem:

Within the insurance sector, claim settlements consist of multiple, semi-structured documents with arbitrary layouts. To ensure proper reimbursements and to prevent fraud, these documents need to be checked for completeness and formal requirements. Preparatory data extraction is time-intensive and currently requires a lot of manual work. To remain competitive, enterprises need reliable, fast and efficient automated document processing. Businesses in the insurance sector, including insurance companies, brokers and online service providers, process 85% of their claims manually. With 7.2 billion claims and 5.2 billion contract documents processed annually worldwide (5.8 billion contracts and claims in the EU), the sector could save an estimated €73 billion in processing costs worldwide by efficient automation of these processes.


Importance for society:

Successful market launch of omni:us will be the first step in a wider disruptive process that has been recognised by the European Commission, i.e. the digitisation of manual document processing. Qidenus wants to be the global leader in this field and aims to set the bar high.

Implementation of omni:us is expected to:

- Replace current manual labour by a digitisation solution with domain focused use case process extensions. By pushing the boundaries of AI, we aim to redefine and improve the work environment. Through the radical substitution of repetitive processes, we will free employees to unleash their full potential, granting freedom to engage in more productive and challenging tasks.

- Increase the competitiveness of EU SMEs in the insurance domain due to more time-efficient processing of documents, enabling the reduction of the time and size of contract and claim handling categories. More cost-efficient processing is expected to result in more competitive execution of services and products.

- Open up new development opportunities for European software enterprises and developers, as further implementation of omni:us will result in customisation/optimisation of individual document processing steps.

- Promote Technology Transfer: omni:us is a direct result from European Technology Transfer of academic achievements (our R&D collaboration with CVC, Barcelona and UPV, Valencia) into the full commercial uptake on a European, and later, a global scale.


Overall Objectives:

The overall objective of the project is, that omni:us will provide enterprises with automated document processing software to reduce the cost and increase the efficiency, speed and quality of data extraction and archiving compared to existing manual and digital methods. Therefore, we plan to further upscale our TRL6 prototype as a B2B SaaS solution for the insurance industry.

The omni:us SaaS architecture will:

1.) Provide our customers with the software tool to improve their processes and turnover;

2.) Provide a platform on which 3rd party developers can offer novel features, extensions and add-ons.
During our project we:

- Performed overall Project Management to ensure that project objectives are met according to the project plans. We also performed intellectual property (IP) management to identify, protect, and exploit the project IP; and knowledge management (KM) activities to gather and assimilate the project know-how.

- Carried out Customer Development activities to validate our business models. We created a Customer Development Plan and a Preliminary Commercialisation plan (strategy, market uptake and replication) and a financing plan for the market launch.

- Carried out Communication and Dissemination activities to support exploitation, targeted towards potential customers, partners in the value chain, potential investors, end-users, and other stakeholders from research, commercial, social, environmental, policymaking, and standardization, educational training bodies etc. We set up a Project Webpage and created a Communication Plan.

- Designed our system specifications and implemented the core framework for the omni:us SaaS platform.

- Defined the technological module features and self-learning algorithms that needed further extended functionalities and optimization.

- Defined objectives for the vertical-adapted User Interface and segment-specific extensions.

- Further developed our technology for the three generic modules (Classification, Recognition/Extraction, and Mapping) to optimize extraction and archiving performance of omni:us.

- We began developing our user-friendly deep domain API SaaS-environment specifically designed for all stakeholders in the Insurance Contract and Claim Management market.

- We prepared a demo that has been shown to potential pilot participants to convince them about our technology and omni:us SaaS.

- We gained 2 pilot customers to perform our pilots to upscale our solution from TRL6 - TRL9 during the second year of the project.

- We created a Business Innovation plan including (strategy, market uptake and replication) and a financing plan for the market launch.

- Successfully developed our technology for the three generic modules (Classification, Recognition/Extraction, and Mapping) to optimize extraction and archiving performance of omni:us.

- Successfully developed our user-friendly deep domain API SaaS-environment specifically designed for all stakeholders in the Insurance Contract and Claim Management market.

- We executed 3 pilot use cases with 2 pilot customers to upscale our solution from TRL7 - TRL9 during the second year of the project.
Progress beyond the state of the art:

The main difference between the SoA and us is, that our technology comprises three processes (classification, extraction and mapping) and unites a combination of computer vision (OCR), handwritten text recognition (HTR) and natural language processing (NLP). We are applying deep process domain (AI) trained document API-extractors focusing on repeating use-cases in the whole insurance industry, which are easily and rapidly implementable to current legacy and workflow solutions.


Expected results by the end of the project:

The project resulted in the market-ready omni:us deep domain API SaaS-product to be commercially launched in the segment of Claim Management.


Potential Impacts:

The social impact of omni:us is two-fold:

1.) By replacing manual editorial labour, omni:us frees-up employee time for more productive and challenging tasks, to create more a higher degree of satisfaction among personnel.

2.) Often, optimisation and digitisation come at the expense of personnel, as to reduce on staff costs. Large firms have the power to move large amounts of labour off-shore, while SMEs due to their multi-utilisation structure typically keep their staff locally and therefore are less profitable. omni:us achieves higher degrees of efficiency especially amongst SMEs, which makes them stronger and more competitive.

Furthermore, omni:us will have a positive environmental impact, thanks to the digitisation and automation of current paper-based processes, reducing the amount of paper being used as more content is produced directly in digital, rather than paper form.
omni:us logo