SAP and Artificial Intelligence

Introduction

SAP and AI a highlight now!
SAP is the leader of ERP market, covering many areas, from accounting, to production management, logistics, etc. So many areas where artificial intelligence has more than its place and can make huge improvements.
If SAP is already working on the subject with its platform Leonardo, everything is already possible TODAY whatever the version of SAP. The purpose of this article is to explain why artificial intelligence on SAP is now
As always if you appreciate the content, thank you for sharing and likes ... This increases the visibility of the article and my chances of doing AI projects in SAP ;-)
For that I will address several points:
  • Why SAP is the best platform for AI
  • Some examples of business cases
  • Technical Architecture and NoAI (Not Only Artificial Intelligence) Components
  • The implementation of IA projects on SAP

Why SAP and AI ?

We can agree on these aspects :
  • SAP is the leader in ERP, mainly for medium and large companies.
  • SAP has huge amounts of data, far enough to drive AI
  • SAP covers many features that can be greatly increased by the AI
  • I often say that what we build with AI is an intelligent entity: the soul the business. Because SAP is often the core application of the enterprise, SAP is the best platform for doing AI

Examples

Here is a list of some examples of AI application on SAP:
  • FI / CO: detection of inconsistency on the financial movements, and imputations, and in real time self-auditing (Machine Learning, Pattern reco, scoring, data science, ...)
  • NetWeaver / DMS: Automatic creation of knowledge bases from all documentations contained in all SAP content servers for texts as well as for images (NLP, Machine learning, big data, artificial vision)
  • MM: Prediction of stock changes and stock requirements and negotiation process in the field of purchases and sales (machine learning)
  • WM: Organizational Optimization and material Distribution (Machine Learning, and Grouped Genetic Algorithm)
  • HR: Analysis and prediction of seasonal phenomena such as diseases to do prevention, stress analysis, career monitoring and anticipation of resignation. (Machine Learning)
  • BC: Business Workflow optimization by automating parts with machine learning (Machine learning, NLP, mathematical scoring to reinforce decisions, game theory)
But also, optimization of manufacturing and transport flows, monitoring of suppliers and competitors (web crawler), .... etc ....
SAP is made for AI, SAP is the core application of many companies! The AI associated with the human intelligence, it is "the magic potion" of the companies, the increased company
What we must understand is that the AI applies especially to:
  • Make better predictions
  • Make better classifications and detect anomalies
  • Automatically search for new strategies
  • Transform texts (natural language) into actionable and intelligible data
  • Optimize the functioning of its SAP infrastructure
  • Etc.
AI to allow SAP to learn by itself

Technical architecture

AI in SAP, using ABAP, the key of a success
Video (French, cut the sound, it should be clear enough)
SAP can connect to everything as a client (service consumer) or server (service provider), but I have a preference for an ABAP friendly architecture
My approach is to create a proxy server ie a black box that will be consumed by SAP in ABAP essentially.
AI and NoAI (not only artificial intelligence) functions are written in the black box. In SAP, via SM59 or connects the server proxy. And in ABAP, as much in functions, methods or in via ALE, we consume in ABAP the methods no AI provided by the black box
In this case we offer the following features:
  • Classify Tensorflow (ZTENSORFLOW) and other behaviors provided by Tensorflow
  • Integration with ElasticSearch (ZELASTIC)
  • Integration with HADOOP components (ZHADOOP)
  • Integration with the Sopra Steria platform (ZSSG) including scoring (DEA, Prometheus 2), fuzzy logic, linear regression, mounted-carlo, data science.
  • Etc.
Our approach is based on a bet: To avoid any change for the SAP functional and technical consultants. They consume features improved by the AI without even asking the question of the AI. We deliver NoAI (Not Only Artificial Intelligence) business functionalities only :-)

How to realize an IA project for SAP

small projects, business features, and measurable ROI!
We can distinguish two tracks:
  • SAP standard approach (functional analysis, development, etc ...)
  • The development of NoAI (Not Only artificial intelligence) proxy server
To initiate a first project there are some steps that we propose:
  • Real demonstrations (not on powerpoint) of different AI functionalities usable in SAP and introduction to the AI for the business and the executives.
  • Demonstrations of AI consumption using ABAP
  • Assessment to capture the individual and collective intelligence of teams wishing to integrate AI
  • Realization of the necessary components in the proxy server (black box)
  • SAP ABAP classic development ....
The good news is that most components are reusable for many needs and that the biggest return on investment in AI is earned with generic components in an average 100-day project.
Whatever the version, SAP is the ideal platform for doing AI. Technically it's easy, and S / 4 favors it even more. The components we offer are:
  • IBM Watson (NLP)
  • Tensorflow (classification, vision, scoring, etc ...)
  • Stanford (NLP)
  • NVidia CUDA (Hyper-parallel computing for forecasting on deterministic models and simulation)
  • ElasticSearch (search engine and NLP)
  • And tailor-made for the predictive, the clusters for the unsupervised training, etc ...

Conclusion

SAP is a complete platform, with huge amounts of data, and is the core application of many enterprises.
The AI is mature, the essential having been created for more than 20 years, what is new is its industrialization thanks to the enormous computation power current, and Big Data.
If the essential project of AI in the company is to improve all business flows (optimization) and ultimately create an intelligent entity that would be the company, SAP should be the priority platform for AI projects
For more information you can contact me at jerome.fortias@soprasteria.com
J

Commentaires

Posts les plus consultés de ce blog

Quand Google fait un KIT AI hyper débutant !!!

50 niveaux de RaspBerry Pi

Netflix AlphaGo - le documentaire à voir absolument