For the longest time, artificial intelligence (AI) was nothing more than exciting material for science fiction novels and movies. In 1982, Harrison Ford was hunting down replicants in ‘Blade Runners’; in 2001, Steven Spielberg wove an intricate tale of a relationship between human and machine in ‘A. I.’
Technologies that emulate or surpass human skills and characteristics are fascinating – not only on the big screen. For decades, scientists have dedicated time and money to the topic of artificial intelligence. The workshop ‘Dartmouth Summer Research Project on Artificial Intelligence’, held in 1956 at Dartmouth College in New Hampshire, USA, is seen as the starting point of AI as academic subject.
Progress in the field of AI wasn’t exactly on the curriculum during that workshop at Dartmouth. However, over the past few years, the futuristic visions of scientists and Hollywood directors alike have become more realistic as the power of computers continues to increase, more digital, connected products are flooding the market, and our society and economy are producing enormous amounts of digital data.
What AI can and can’t do
What we are missing is a valid common definition of AI or, more generally, what counts as intelligence. What we can say without the shadow of a doubt is this: To some degree, machines are capable of autonomously making decisions and acting on them. They are capable of recognizing and analyzing patterns in data, language or images. In most of these cases, machines do significantly better than humans, as their analysis is usually faster, more comprehensive, and more accurate. Technology-based methods – like machine learning – are a requirement for these processes to work seamlessly.
All of this falls under the category of ‘weak AI’, meaning that a software can autonomously solve specific problems. While the software takes care of the task that it was programmed for, it will never truly understand it nor form what we might call a conscience – which are the required criteria for ‘strong AI’. Some scientists expect ‘strong AI’ to develop superintelligence with (simulated) conscience that will surpass any human and be far beyond our control in the future. As of now, this is more on the science-fiction side of AI, though.
AI: Potential value of €430 bil.
Since ‘strong AI’ is still a distant vision of the future, today’s focus is more on ‘weak AI’, meaning manageable, ethical software and solutions for practical use cases. According to consulting firm PwC, (weak) AI alone will generate value of €430 billion in Germany until 2030. In another study, PwC studied the most common use cases of artificial intelligence. A majority of respondents indicated that they used AI for automating existing business process and analyzing data. AI is the most useful when dealing with one individual, complex task involving an enormous amount of data.
Obviously, AI has enormous potential for almost every company. SAP customers are therefore faced with the question of which technologies SAP offers them to upgrade and enhance their existing system landscapes. There are some applications in the ERP company’s portfolio that sound like AI.
Leonardo: Approximation to AI
Innovation platform Leonardo is the most prominent example. SAP introduced it in 2017 with all the fanfare of a groundbreaking revelation, but it has gotten eerily quiet around the solution portfolio. Leonardo comprises different applications and microservices to a variety of hype topics, like machine learning, IoT and blockchain. It was an ambitious strategy, but SAP seems to have given up on it. Instead of being seen as all-compassing technology provider, SAP seems to have gone back to its roots as practical solution provider. SAP Leonardo has been rebranded as Intelligent Technologies, but many of the technological and functional services it once offered have been or will be discontinued in the foreseeable future.
Automation through ML and AI
Instead, SAP is now focusing on business processes and company-specific requirements by integrating innovations and individual solutions. SAP offers customers different ways to join the world of artificial intelligence, each of which is tailored to different target groups while still complementing the others.
In its new ERP system S/4 Hana, SAP has embedded some AI functionalities which are supposed to make individual S/4 applications intelligent. The goal is the automation of routine tasks to relieve employees of monotonous work, shifting their focus to more complex scenarios. There is more than one approach to reach this goal.
- Pre-defined scenarios. SAP offers pre-defined scenarios which have to be trained with the client company’s historical data before they can automate routine tasks. This is most beneficial for e.g. assigning incoming payments to open invoices or creating purchase orders.
- Robotic Process Automation. Robotic Process Automation offers users the possibility to create custom AI-based, automated processes to have bots take care of recurring tasks by using familiar patterns.
- Situation handling. By leveraging pre-defined scenarios and Robotic Process Automation, situation handling will be a focal point: As recurring tasks have been automated, users will only need to react to unusual occurrences in the system.
- SAP Conversational AI. As part of Leonardo, SAP Conversational AI was meant to realize chatbots. As part of S/4, however, it enables interaction with the system via natural language.
SAP Analytics Cloud
As strategic analytics solution of the future, the SAP Analytics Cloud (SAC) not only includes common BI functionalities, but also AI-based predictive functionalities called SAP Smart Predict. The tools are designed for Citizen Data Scientists, meaning ambitious Business Users. The Smart Predict applications offer everything Business Users might need for predictive analysis, e.g. to predict the most suitable time for offering discounts. The necessary AI models can be trained with historical data.
SAP Data Intelligence
Running on SAP Cloud Platform (SCP), SAP Data Intelligence is by far the most elaborate AI solution in SAP’s portfolio. While it is the most likely ‘successor’ of Leonardo, the focus has shifted to business. SAP Data Intelligence combines a more conservative business world with open source, making it more open to different solutions and applications, like Jupyter Notebooks (Project Jupyter), Python and Python-based frameworks for machine learning like pandas, scikit-learn or TensorFlow.
Focused AI strategy
With SAP Data Intelligence, the created data models can be transferred to a company’s SAP landscape for further development, testing, and performance monitoring. And, most importantly, AI scenarios can be audited, complying with regulatory guidelines.
SAP has recognized that AI is an important – if not the most important – technology topic and is enthusiastically adding to its existing portfolio without losing sight of business requirements. Actively researching AI is not SAP’s forte, however. Companies like Facebook, Google or Microsoft as well as open source initiatives are and will continue to be the driving force in AI innovation.
If companies want to implement AI scenarios, they first should evaluate (after, of course, already having evaluated the profitability and feasibility of the desired use case) what they need and what SAP offers. Where SAP solutions leave something to be desired, SAP partners or third-party solutions can usually help fill the gaps.