The concept of the “intelligent enterprise” was the centrepiece of SAP co-founder Hasso Plattner’s keynote at this year’s Sapphire conference in Orlando, Florida.
Plattner, and other leaders at the supplier – notably Robert Enslin, president of cloud business group SAP – said the intelligent enterprise is not something customers can order off a price list.
Instead, it is the way they want to express SAP’s goal to raise its fundamental offer of financial applications and materials resource planning to one of making the so-called digital transformation of companies and other organisations possible.
For the past eight years, the supplier has put its in-memory, columnar database Hana at the core of its offer to customers.
That expanded to S/4 Hana, its ERP [enterprise resource planning] in February 2015, and, at this year’s Sapphire, C/4 Hana, an umbrella term covering its various customer relationship management (CRM) products and services.
Last year, it announced Leonardo, a so-called “digital tool box” that includes the supplier’s machine learning efforts, blockchain services and internet of things activities.
At this year’s conference, SAP announced a slew of “partnerships” and products “to enable enterprises to become more intelligent, with expanded capabilities from advanced technologies such as conversational artificial intelligence, blockchain and analytics”.
ML roller coaster
Markus Noga, senior vice-president, of machine learning at SAP, said the supplier’s strategy is focused on its traditional strengths in enterprise software and taking an ethical approach.
Noga, a former Booz Allen Hamilton strategy consultant with a computer science doctorate from Karlsruhe Institute of Technology, is head of machine learning for SAP globally.
“It’s been a roller-coaster ride. It’s exciting, but carries obligations, too. All product units are platform teams looking to add AI. And there is customer interest along so many different dimensions.
“There are two categories we have to say ‘no’ to. One is [the category of] things we should not be doing, that are not in line with our values. The other is that which is outside our capacity or priority. There is a whole host of industries where we are banking on partners”.
SAP’s distinctive strategy with respect to machine learning is shaped by a guiding vision of the intelligent enterprise. “We want to make business software self-driving and autonomous, where repetitive functions are automated, and where it is easy to consume by a human, because you can interact with it using natural language,” he said. “And also augmenting human intelligence, where the human touch is required.
“We are not focused on building the best compilers to translate [neural] networks to work on the best hardware. Nor are we focused on being a hyper-scale provider of cloud infrastructure. That’s why SAP Cloud Platform follows an open, multi-cloud strategy and runs atop the major infrastructure providers,” said Noga.
“It’s why we are working with open source providers of machine learning frameworks, like [Google’s] TensorFlow. And why we are partnering with Nvidia, with their GPU hardware that is providing so much of the deep learning research and projects in the world now”.
SAP has also set up an AI Ethics and Society Steering Committee, which Noga calls to attention.
“SAP’s overall mission is to help the world run better and improve people’s lives – and not all uses of this technology do that.
“For example, you could create the perfect [digital] sales assistant that coaches you, but you could also create the perfect sales policeman that monitors your behaviour, and if you don’t do the optimal thing, calls in your supervisor, who raps you on the wrist,” he said. “The ML model we would have to build for both these cases is exactly the same. One makes the world a better place, and the other makes you miserable”.
SAP Cloud Platform – public and private
The supplier also announced that its SAP Cloud Platform, known as SCP, has been made available on Google Cloud Platform and Microsoft Azure, alongside Amazon Web Services (AWS).
The supplier also intends to release SAP Cloud Platform, private edition, on IBM Cloud as a private cloud deployment. This is aimed at customers in regulated industries such as banking, healthcare and transport, as well as those managing sensitive data, to have – it is said – the “flexibility, speed and agility to innovate without jeopardising security and control”.
Bradley Knapp, global offering manager, SAP certified infrastructure, IBM Cloud, also spoke to Computer Weekly at the show. “Lots of our customers who operate in regulated industries and the public sector want the flexibility and the broad development environments [of the SAP Cloud Platform] but also want, but a multi-tenant public cloud model is something they just can’t do,” he said.
“So, this edition runs in a dedicated stack in a virtual private cloud in IBM cloud, where SAP holds the contract with the customer. IBM then provides the entire stack of services from infrastructure level all the way up to the run time”.
The edition will be generally available from Q3 of this year in North America. “It’s another example of how IBM is trying to design our cloud to make it easy for large business to come in, and get the advantages of cloud,” said Knapp.
Elements of ML
The elements of the SAP machine learning product set unveiled at Sapphire included:
SAP Conversational AI, which enables companies to develop chatbots which can be integrated with SAP and non-SAP systems and are available as preconfigured industry-specific bots. France’s railway company SNCF and telecommunications provider SFR are using SAP Conversational AI, the supplier reported.
SAP Leonardo Machine Learning Foundation, which allows customers to develop individual applications, has five new services, including object detection, text recognition in images and text classification, which analyses and automatically categorises text documents.
SAP Cloud Platform Blockchain as a service enabling enterprises to build and extend business systems with blockchain technologies, such as Hyperledger Fabric and MultiChain. The supplier said 65 companies participate in the SAP blockchain co-innovation initiative in manufacturing and supply-chain activities. US sausage maker Johnsonville was given as an example of a company that traces the origin of products across its supply chain.
SAP has also started a global blockchain consortium with seven founding members, including Intel, Hewlett Packard Enterprise and A3 by Airbus.