Neural Network Mind Reduced_1

FIS / FIS EDC Machine Learning

00SAP Invoice ProcessingTags: , ,

 

Artificial Intelligence and Machine Learning Introduction and Developments

Find out how FIS and SAP are creating new machine learning technologies to solve real business problems

Neural Network and machine learning

SAP Business Technology Platform

Artificial intelligence (AI) and machine learning (ML) are now being used by organisations of all sizes and across a diverse range of industries. Intelligent technologies support the digital transformation of companies and are becoming more important. 

SAP provides basic functionalities for the development of modern Artificial Intelligence solutions with the SAP Business Technology Platform. The portfolio includes numerous innovative technologies, such as Artificial Intelligence, Machine Learning, Internet of Things (IoT), Big Data and Blockchain.

The platform provides application developers with comprehensive functions and a wide development environment they can use to enhance, integrate and quickly create innovative applications themselves without any infrastructure maintenance tasks.

Your benefits by implementing intelligent SAP Technologies

  • Smooth integration of AI applications
  • Easy adjustment, enhancement, optimisation and automation of processes
  • Simplified and faster work flows
  • Data-driven services can be provided immediately
  • Detailed forecasts and, consequently, faster decision making on a strong data basis
  • Creation of new data-based business models
  • Improved adaptability to market changes
  • Increased competitiveness and efficiency
  • Directly measurable added value through cost reduction and time saving

Artificial intelligence and Machine Learning

Artificial Intelligence means the capacity of machines and systems to imitate human actions in order to simulate intelligent behaviour. It is either the fulfilment of specific tasks, such as advertising personalisation (applied AI), or more comprehensive activities, where entire systems or devices completely imitate human behaviour (general AI).

Contrary to many opinions, Machine Learning (ML) is not the same as AI. Machine Learning is part of Artificial Intelligence focusing on algorithms and mathematic models.

This intelligence is “trained” in such a way that it can solve any occurring problems by itself by developing the appropriate AI algorithms. The bigger the dataset and the number of sources, the more successful AI technology will be.

Machine Learning is essentially a sub-division of Artificial Intelligence. It is about imparting knowledge to machines or computers. To this end, a neural network or models need to be trained with data, such as texts or images.

The respective data can be unstructured financial data from the SAP system for instance. From the mass of training data, the system extracts characteristic features to enable a direct recognition and assignment of the data imported in the future based on these features.

Overview of SAP intelligent technologies

  • Conversational Artificial Intelligence (AI or CI) includes the entirety of intelligent technologies standing behind automated message transmission and voice-activated applications. They enable human-like interactions between computers and humans. As a result, users will be guided and supported in their daily business.
  • Intelligent Robotic Process Automation (IRPA) is the combination of RPA, ML and AI. Tasks, such as the processing of messages and orders or the execution of data transfers, are assigned to digital robots or “bots”. These bots efficiently implement time-consuming and repeating processes.
  • Internet of Things (IoT) means the networking of devices, machines and installations with and through the Internet. Sensor data transferred in real time increase the transparency of business processes. Decisions based on this data will be made proactively and more quickly.
  • Machine Learning (ML)  IT systems (machine learning models) are trained on the basis of existing datasets and algorithms enabling them to identify patterns and rules and use them to develop results by themselves. In this way, business decisions are optimized or even made automatically.

FIS Machine Learning Applications and Development

Usually, self-learning technologies, such as ML, become relevant if applications cannot be optimised or developed in a rules-based way. The FIS professionals work every day on testing and further developing innovations by using ML, RPA and bots. FIS, for instance, enhances its own products through the use of machine learning or further develops customer-specific ideas and proposals.

FIS strategy has always been to put the customer requirements at the centre of our thinking. That is why FIS offers both SAP applications and our own self-developed optimisation applications, with teams of experts who will work with you to find technology solutions for your specific challenges.

As an example, within our SAP integrated invoice processing solution, FIS / edc (invoice monitor), the two main technology components use machine learning technologies to automate the invoice throughputs with less human intervention.

In the FIS OCR data reading application (FIS / fci), the embedded machine learning technology is able to extract the key content from an invoice in the same way that an individual would recognise the data (i.e. data in the vicinity of key words and prefixes) and then later stores the key data in a knowledge base for subsequent invoices from the same supplier.

In the SAP integrated invoice processing module (FIS / edc), machine learning can be used to create automatic account assignments and direct FI invoices to the correct individuals or groups based on the experience of previous postings.

 

 
FIS EDC invoice processing diagram

00SAP Invoice ProcessingTags: , , ,

OCR invoice data collection technologies and functionality

Advanced technologies work together ensuring that invoice data is collected automatically and accurately. The invoice reader uses validation and viability checks prior to the data export to SAP or Financials.

Introduction
The video highlights some of the advanced technologies that are applied to the task of automated invoice data collection.
The technologies ensure that invoices are correctly classified and the data quality is maintained, prior to transfer to SAP / ERP.
The technologies significantly reduce the workload of the accounts payable team.
Who is it for?
For organisations with SAP ERP (ECC 6 or S/4 HANA) who wish to:
  • Replace time consuming tasks such as manual data-entry when processing invoices
  • Improve the invoice data quality in SAP  by utilising tables from SAP for company code identification, PO validation, current tax values, and supplier identification
  • Make savings by reducing invoice processing costs, making a significant improvement in data quality
  • Reduce the workload for the accounts payable team
  • Reduce the payment times for vendors by processing invoices in a timely manner
Key Information:
Functions of OCR Reader and Integrated Invoice Processing
 

Emailed invoices are automatically forwarded to FIS / fci (the invoice data collection component) and then archived for retrieval in SAP / ERP. Therefore, no need for manual tasks such as printing, scanning and keying-in invoice data

Multiple OCR engines are employed. The data from all engines is compared to support data quality.
Invoice classification is confirmed against standard tables in SAP / ERP to highlight potential errors in the invoice data.
"Fuzzy logic" is used for comparing invoice data (i.e. supplier recognition) with reference data to compensate for inexact data on the invoice or within the SAP master data.
Freeform recognition technology automatically searches for data in the vicinity of keywords to collect the key invoice data from day-one.
Machine learning. An intelligent knowledge base is used to collect the key invoice data. The knowledge-base uses machine learning by analysing the co-ordinates of previous successfully read invoices from each supplier.
 

 
schematic diagram working from home with FIS

00SAP Invoice ProcessingTags: , , ,

 

Automated invoice data collection from day one

Advanced OCR technologies are employed to accurately collect the key data from invoices and other financial documents. The invoice data is checked for viability prior to posting into SAP or other ERP/Financials. As most invoices are processed without manual data-entry, workload is reduced, financial savings are generated and data accuracy improved. 

Introduction
By introducing technology to automate invoice data collection and check for potential errors, the invoice reading process becomes more efficient and the quality of the key invoice data is improved.
Accounts Payable Teams will benefit from  reduced workload as the number of data-entry tasks and reduced and the pre-SAP/ERP classification of invoices is eliminated.
Who is it for?
For organisations who wish to
  • Automate invoice data collection and reduce the workload of their accounts payable team
  • Remove labour intensive and expensive stages of invoice classification by eliminating the need for bar-code label additions to invoices
  • Make savings by capturing invoice data at the first stage (early invoicing) rather than scanning invoices (or using a bureau service) retrospectively
  • Support the accounts payable team working from home by recommending that suppliers send invoices electronically via email attachments
  • Improve data quality by installing technologies with validation and viability checks
Automating invoice data collection and reducing manual data-entry
The FIS invoice reading technology is designed to model the way in which a human would inspect an invoice – reviewing the key invoice data (using multiple OCR engines) and searching for the data in the vicinity of key words and phrases.
Data is validated and against SAP/ERP tables where applicable. 
The data from each supplier is stored automatically in a knowledge-base that gains experience with each supplier invoice.
Tasks that are replaced include
  • Significant reduction in workload as data-entry tasks by the accounts payable team are automated
  • No requirement for bar-code labels to be added to each invoice
  • No requirement for emailed invoices to be printed or scanned
  • No retrospective scanning of invoices or using a bureau service for "late archiving"
  • No need for manual invoice classification into categories such as company codes, currencies, invoice type (PO or PO-exempt)
Remote and hybrid working

Invoice processing remote working

The invoice data collection is an ideal technology for accounts teams that work remotely. The technology works automatically with emailed invoices where they can be forwarded directly into the OCR  technology, FIS / FCI, for automated invoice data collection.

 

 

 
reconciliation balances

Automated Statement Reconciliations

00HomeTop, Products, SAP Statement ProcessingTags: , ,

Automated Statement Reconciliations for SAP 

The FIS monitor provides automated statement reconciliations between vendor balances in SAP and the vendor statements. The SAP integrated application may also be used to reconcile payment advice notes from customers.

balances illustrating automated statement reconciliation

FIS / edc (statement monitor) is an integrated SAP technology that reduces the workload  when reconciling vendor statements or customer payment advice notes (remittances).

The automated statement reconciliation process can be then be achieved in less time and more economically than with SAP alone.

By freeing up time and reducing workload, the Accounts teams can then focus only on the invoices discrepancies between the statement data and SAP vendor data.

Releases of the technology run with both SAP ECC 6 and SAP S/4HANA. The application is able to process both vendor statements and customer payment advice notes in a single SAP add-in technology.

The integrated software in normally paired with the FIS OCR reading technology. This key data reading application automates the  data entry of key data from the statements. Statements can be also received via email attachments.

Automated Statements Reconciliations for Vendors

Vendor Statements usually take a lot of time to reconcile manually. In most instances - all statements are compared with the vendor account balance on SAP.

The statement monitor is focused on identifying vendors with unequal reconciliation balances compared to SAP. The solution’s main advantages include:

  • Reduced workload on the Accounts Payable team, who now only review vendor accounts with unequal balances
  • Ability to identify specific invoices from vendors that have been received or could have been misplaced
  • Improving the efficiency of the accounts payable team - who now only review reconciliation exceptions identified by the statement monitor
  • Statement data is automatically compared with the vendor balances in SAP on the date of the statement

Summary

For organisations who wish to make savings by introducing an automated statement reconciliation  process of reviewing vendor statement balances.

The statement monitor filters exceptions - statements that do not reconcile against vendor balances in SAP on the date displayed on the statement.

An alternative to

  • Manually reconciling statements against SAP vendor records
  • Spending time identifying which invoices from the statement are not in agreement with SAP

Highlights and key benefits:

  • Automatic OCR reading of vendor reconciliation contents from an emailed attachment or a scanned document 
  • Automatic comparison of data from the vendor data within SAP, filtering statements with non-zero balances
  • Identifies invoices potentially in dispute or lost invoices
  • Manual comparison of data now diminished, account team now directed to reconciliations where assistance is required
  • Vendor data from the document is compared with data in SAP in familiar a SAP GUI

Automated Payment Advice Note Reconciliations 

The document monitor can also be utilised to increase the level of automation and decrease the level of manual tasks required when reconciling customer payment advice notes (remittance advice notes).

Using the monitor, customer remittances can be matched and subtracted automatically from the outstanding invoices records in SAP prior to the payment reconciliations with the bank.

Therefore the automated process creates savings and improves matching accuracy.

Highlights and key benefits

  • Identifies invoices sent to customers that have not been acknowledged
  • Improves the efficiency of the accounts receivable team - who now only review reconciliation exceptions
  • Fewer manual comparisons of data required 
  • Customer data from the document is compared with data in SAP within a familiar SAP environment GUI  and therefore easier and quicker to adopt

Automated Data Entry from Day One

calendar day one

Automated Invoice Content Extraction

00HomeTop, Products, SAP Invoice ProcessingTags: , , ,

Automated Invoice Data Collection

Automated invoice data collection from day one. The OCR (optical character recognition) technology extracts the key information from invoices irrespective of their layout. Advanced machine learning technology is applied for improved recognition rates

Artificial Intelligence and Machine Learning - automated data entry

Automated invoice data collection using the FIS OCR document reader

The FIS OCR document reading software (FIS / fci) is a world leading, innovative technology that is used to automatically examine a digital invoice (or other financial document) and collect the required key data with minimal or no human intervention.

The technology generates financial and time savings by reducing the manual workload of accounts payable departments. Data quality is also increased prior to importing it into SAP or other ERP/Financials software.

Using the technology results in less data entry and therefore frees up resource for accounts departments.

The technology collects the content of invoices that are captured as a digital scan or are received as an emailed attachment.

The software offers unmatched recognition rates for invoices, extracting both the header and line-item (position) content.

The OCR technology is normally paired with the FIS / edc (invoice monitor) to post a high percentage of invoices automatically into SAP (ECC 6 or S4HANA) without the need for human intervention (no touch processing).

Data collection on day one

The OCR reading software is pre-programmed to identify specific information fields on an invoice and then extract the relevant data within close vicinity. The ‘freeform’ recognition technology automatically finds the invoice data irrespective of invoice design and layout.

  • Supplier details and invoice receiver checks are validated against the latest SAP data.
  • Invoice financial values are checked for feasibility by checking arithmetic agreement between net, tax and gross values.
  • Fuzzy technology compensates for incomplete or inaccurate data or poor quality invoice scans.
  • Several OCR engines are utilised to increase data reading accuracy.
  • Machine learning, any manual confirmations or checks are stored in an intelligent knowledge base for each vendor to build up a statistical probability of the key data locations.
  • Automated data entry means that the cost of processing invoices is reduced and savings are generated by the technology.
Reporting

Detailed reports illustrate vendor invoice volumes and data extraction rates for specific fields to identify issues with suppliers.

FCI report recognition rates - automated data entry

Summary

For organisations who wish to

Make savings by introducing automated data entry processes and significantly reducing the workload on accounts departments.

Improve data quality by introducing automated validation and feasibility checks  against SAP data.

Alternative to
  • The printing and manually keying in data from a physical document
  • The slow process of data-entry of  invoice data with inherent errors
Key Technologies
  • Multiple OCR engines - working in agreement
  • Automatic recognition of suppliers,  company codes and order numbers
  • Automatic forwarding of emailed invoices
  • Machine learning by storing key invoice data from previous invoice completions
  • Freeform recognition, FIS / fci is pre-conditioned out-of-the box to search for key financial document data
  • Intelligent knowledge base which stores data co-ordinates for all successfully read invoices and therefore reduces the number of corrections in the future
  • FIS / fci collects header and line items automatically
  • Contributes to no-touch processing when combined with FIS / edc (invoice monitor)
  • Ability to read any semi-structured document such as sales orders, payment advice notes or vendor statements