KPMG Document Reader

Document Reader Brochure

查看

KPMG Document Reader applies artificial intelligence according to business scenarios defined based on our advisory experience to explore data value. It helps enterprises create an intelligent system capable of reading documents more effectively and efficiently than their traditional document processing systems, and at the same time facilitating the business scenarios. Harnessing Python, optical character recognition (OCR), natural language processing (NLP) and other technologies, it is able to structurally parse and compare image, PDF, and other files based on flexible extraction and verification rules, and to enable AI-empowered services in tailored business scenarios.

产品特点

Intelligent extraction

It can extract critical information from documents, fill in forms, and verify content consistency, thereby making reviews easier. It supports more than 80 types of document, including contracts, bond issuing documents, prospectuses, auditors’ reports, civil judgement documents, and personnel change announcements; and can label different documents and extract user-defined critical information in complex scenarios with an accuracy exceeding 90%.

Intelligent verification

Intelligent verification can be performed before manual reviews using information extraction models trained with a huge amount of sample data, so as to significantly improve the efficiency and accuracy of manual reviews. It is able to check capitalisation consistency, completeness, compliance, terms, etc. of financial information, formats, and disclosures. It enables one-stop verification of multiple documents using different flexible verification rules for different scenarios in sectors such as finance, banking, manufacturing, pharmaceutical, communications, legal, audit, media, and government.

Intelligent comparison

Leveraging flexibly configured rules, it can compare documents, present the differences in an intuitive manner, and export accurate and searchable results. More specifically, it is able to compare characters, sections, punctuations, paragraphs, semantics, financial data, contractual terms, bond ratings, etc., and identify file differences using different colours to enable quick location of such differences. It supports more than 10 types of document, including contracts, annual reports, prospectuses, bond issuing documents, auditors’ reports, and fiduciary management agreements.

应用场景

Financial sector

Reviewing and filing contracts: This platform leverages professional OCR and extraction technologies to structurally process standard certificates, tables, and contracts prepared in the form of an image or PDF file, and automatically inputs the information extracted. Requesting services and verifying documents: This platform can directly interconnect with enterprises’ internal business systems and risk review systems to automatically review and archive documents. In case of a risk alert, the case will be transferred for manual processing. Parsing and comparing financial statements: This platform can extract and compare critical information from financial statements of enterprises, making it easier to analyse their operating performance and other capabilities.

Automotive sector

Requesting services and verifying documents: When handling finance leasing requests, granting retail loans, and filing documents, enterprises need to manually review a large number of scanned copies, which requires much time and effort. KPMG Document Reader can serve as a platform that harnesses the OCR technology to extract data from the scanned copies, facilitating intelligent business reviews. It is able to integrate business system data to automatically compare and label business data in the scanned copies and synchronise the results to business systems. Comparing bidding documents for duplicates: During bidding, many documents need to be submitted by bidders to project owners, and some of them many be fraudulent. This platform can compare the documents to quickly identify unexpected duplicates, helping with preliminary screening.

Pharmaceutical sector

Reviewing and filing contracts: Some enterprises need to host workshops frequently, which will involve contracting with lecturers and suppliers. This platform can leverages professional OCR and extraction technologies to structurally process standard contracts and attendance registers prepared in the form of an image or PDF file. Categorising drug documents and filing applications: Some enterprises need to apply for permits for selling drug.This platform directly interconnects the internal business systems of the enterprises, so as to categorise and upload the scanned copies by “SKU + document type” to facilitate reviews. The latest version can be viewed on application filing platforms hrough interconnection with customers’ systems. This will help dramatically reduce the labour cost and increase the application filing efficiency.

Government sector

Reviewing expenses: Expense reviews involve a large number of documents, such as bank receipts, business receipts from fund systems, and payment receipts.This platform leverages professional OCR and extraction technologies to identify the types of documents uploaded in batches, extract key information therefrom, and synchronise the information to internal systems of enterprises. It can also implement tailored verification rules to verify the key information, so as to accelerate the business process. Extracting information from standard documents: This plat form can obtain information from VAT invoices, VAT returns, expense claim forms, and various certificates. This will better enable to provide business departments with structured data promptly and accurately, thereby reducing costs.