How SAP is infusing AI into its business applications

SAP’s newly minted chief artificial intelligence (AI) officer, Philipp Herzig, has outlined the company’s AI strategy, which focuses on infusing artificial intelligence capabilities into its business applications in a bid to make AI more accessible to customers.

Speaking to media during a recent visit to Singapore, Herzig, who reports directly to SAP CEO Christian Klein, noted that the foundation of SAP’s AI strategy lies in its ability to bring together data and processes across SAP applications to support what it calls business AI.

“Business AI means, first and foremost, that we are embedding AI into our business applications – supply chain, finance, HR, procurement, travel and expenses – and various other business processes,” he said.

SAP has already started baking AI capabilities into its SuccessFactors HR suite as well as its Sales and Service Cloud. These capabilities, among others, include leveraging generative AI to help recruiters create job descriptions and customer service teams to improve ticket resolution times, for example.

It has also developed the Joule co-pilot, which not only provides contextualised information and assistance for users of SAP applications, but also improves developer productivity by incorporating code-generation capabilities for data models, application logic and test script creation.

“We’ve had a constant drumbeat of new innovations around using Joule to generate job descriptions, and we are now expanding to even more use cases in HR,” said Herzig, adding that SAP has built capabilities for about 30 AI use cases, and these will be expanded to cover over 100 use cases this year.

On how SAP is prioritising AI use cases, Herzig said the company does so based on value: “You can do a lot of with AI, but if the value isn’t there, or if it’s too expensive, it will not get adopted. We always look at it through the lens of the return for customers and whether it’s part of existing licensing or part of our premium AI offerings.”

“Of the 27,000 customers who are actively using SAP business AI today, less than 1% are on-premise. Going forward, this is the only way we’ll design AI, otherwise we will not scale with the number of customers we have”

Just as important is the need to ease AI adoption for customers. Herzig said with embedded AI, SAP has “learned the hard way that if it’s not delivered as a service, and not through cloud and out of the box, adoption will not happen”.

“Of the 27,000 customers who are actively using SAP business AI today, less than 1% are on-premise,” he added, noting that those who consume it as a service have reaped immediate benefits. “Going forward, this is the only way we’ll design AI, otherwise we will not scale with the number of customers we have.”

SAP’s approach also makes AI capabilities more accessible to organisations that may not have the expertise to implement the technology. “If we, for example, ship a model, and then you need to retrain that model and do the data cleansing, those things take a lot of skills, and many customers don’t have 100 data scientists,” said Herzig.

Still, organisations with the skills and that want a slightly different version of an AI-enabled application can tinker with the technology through the SAP Business Technology Platform (BTP), which comes with pre-built integrations across SAP applications.

“Our customers and partners can slightly modify or customise what we have designed to build their own versions of the applications,” said Herzig, adding that they can do so faster through BTP which alleviates integration pain points.

“You can build it on Azure, Google Cloud and Amazon Web Services, but you need to get into the security, build your data pipelines, integrate with identity management, and so on. This is where our strategy makes sense because we needed to solve those challenges ourselves. We are giving BTP almost as a by-product to customers to build their custom applications.”