Business

By now, every second company worldwide uses at least one application powered by artificial intelligence. But large barriers for adoption remain. What are those blockers and how can we innovate to remove them? SAP reveals the priorities for its new AI-focused SAP Innovation Center in Newport Beach and kicks off an innovation campaign with HeroX to accelerate the impact of enterprise AI.

AI has entered our daily lives. We talk to digital assistants, we use smart apps to forecast how we will look in 50 years, and we browse through personalized social media streams based on our preferences and behavior.

Machine learning and AI have also been on the rise in the enterprise, presenting unique challenges to the businesses that embrace these new technologies. SAP develops solutions that assist in all core activities to reduce repetitive work. With SAP Conversational AI, SAP Data Intelligence, and intelligent robotic process automation, SAP has created a multifaceted foundation for infusing intelligence into applications of end-to-end business processes.

However, a recent study by McKinsey reveals that even though every second business is now using at least one AI-powered application, foundational barriers for mainstream use remain. Some of the most cited challenges are about functional silos that constrain end-to-end solutions, lack of available data, or limited usefulness of data, as well as that personal judgment overrides decisions made by AI.

Without questioning the progress of AI, the study clearly shows that there is much more potential. SAP is tackling those challenges to deliver a truly intelligent enterprise. We want to make software systems more human and bring data-driven business and data protection together.
Launching the New AI-Focused SAP Innovation Center in Newport Beach

On September 6, SAP will celebrate the grand opening of the latest location of the SAP Innovation Center Network. In addition to stunning new offices spaces and our HanaHaus concept, some of SAP’s smartest minds are already at work. Under the leadership of Hans-Martin Will, who previously led the development of AI systems for Amazon Alexa, a new engineering team is working on solving some of the toughest problems in enterprise AI, building on the experience of our central machine learning organization at SAP.

The team focuses on:

Ambient voice technology: Fifty-seven percent of IT decision makers whose enterprises are currently investing in voice technology believe accuracy is the biggest challenge. We want to change this and make voice enterprise-ready. And to infuse premier research into our efforts, we are partnering with the University of California, Irvine.

AI-driven development: Artificial intelligence is being embedded into every business function, and software development itself isn’t an exception. We want to make it easier for developers to incorporate AI features into applications as well as create capabilities that intelligently automate parts of the development process itself.

Explainable AI: Often personal judgements still override decisions made by an AI. Why? If a person makes a decision, she usually justifies it, backs it with facts, and lays out arguments to make it comprehensible for everyone involved. A machine doesn’t do that. To create more trust and accelerate the positive impact of AI, we believe it needs new strategies to make AI explainable.

Data labeling and anonymization: Machine learning requires large amounts of training data to be effective and has a high demand for data utility. Typical data anonymization methods are either not sufficient to secure the sensitive information or drastically reduce data utility. We want to identify new data labeling and anonymization techniques that help overcome this challenge.

SAP and HeroX Reward Your Ideas with $60,000

One of our first public initiatives in this latter focus area is a prized competition we are now launching together with crowdsourcing platform HeroX. Starting today, we invite technology experts and innovators all over the globe to contribute their ideas. We challenge the bright minds out there to identify new techniques that allow us to anonymize sensitive information without compromising the results of machine learning. Precisely, we look to improve the accuracy with which we can translate semi-structured information into structured data by encouraging companies to share real business documents without disclosing personally identifying data.

If we can find scalable and transferable techniques for secure data anonymization in semi-structured or unstructured data, we could use all sorts of document, images, or even videos to train machine learning models. This would be a great start for building better AI without worrying about privacy and data protection.

We want to use the power and speed of communities that openly co-innovate to solve this tough problem. What SAP brings to the table are strong guidance, real customer cases, and the broadest and deepest business process expertise you will find in the software industry. The best contributions will be rewarded with $60,000 overall. As sharing creates value, we encourage the innovators to open-source their winning solution for this challenge after the competition has ended.

There are still big challenges in AI that we need to solve, and many of them go beyond just purely technical work; they require enterprise context, a good sense for social responsibility, and sound knowledge about international standards. Our latest initiatives clearly outline the path forward: SAP is committed to unleashing the full power of enterprise AI in the best interest of our customers, delivering both business integrity and ultimate value from data.