Search
Close this search box.

Using AI-Supported Processes in SAP for Document Management

Aicomp’s trainees test SAP Leonardo AI application for automated document management.

SAP Leonardo supports artificial intelligence

SAP combines a solution portfolio for data-driven intelligent technologies under the name “SAP Leonardo”. As a digital innovation system, SAP Leonardo provides applications and microservices for machine learning, IoT, and blockchain to enable the processing and analysis of large amounts of data and to provide companies with new technological scenarios and approaches to solutions and approaches to analysis.

As part of the SAP Leonardo Machine Learning Portfolio, the SAP Leonardo Machine Learning Foundation provides users with intelligent services based on frameworks such as TensorFlow. These services are versatile and cover areas such as speech, image, and video recognition. In addition, sensor data can be evaluated and adjusted. This in turn allows time series and analyses to be created, which then enable predictions to be made using this data.

Aicomp’s trainees test AI application for document management

Aicomp specifically promotes the trainees’ involvement with current IT trends. Artificial intelligence is an exciting future topic for this purpose. In connection with Aicomp’s many years of experience with SAP systems, the consideration of possibilities using SAP Leonardo was obvious.

Automated invoice reconciliation via AI

To start with SAP Leonardo Machine Learning, the SAP Cloud Platform Cockpit is required. The SCP Cockpit offers numerous services. “Document Information Extraction” is one of them. The trainees have set up and tested this service as prototypes. It enables PDF invoices with content in headers and tables to be processed. In doing so, the AI recognizes a pattern and can read out various data such as invoice number, addresses, quantity and much more. This data can be used to create invoice reconciliations. This replaces manual invoice evaluation and saves a lot of time.

AI prototype for the recognition of handwritten data

As one of the largest and best-known providers of AI frameworks, Aicomp trainees have also developed their own AI using TensorFlow. This AI makes it possible to recognize and reproduce handwritten numbers. The AI can also be extended by a letter recognition.

As of version 2.0 SPS02, SAP HANA contains the external machine learning library, which enables the integration of TensorFlow. The External Machine Learning Library (EML) is packaged as an Application Function Library. Models provided by TensorFlow are registered in HANA via remote sources and then called via an SQL script.

More News