Why did SAP’s CTO make scaling a generative AI strategy?

Every technology leader embarking on an AI journey faces an inevitable question: take a limited approach or go all-in?

For Juergen Mueller, SAP’s chief technology officer, the answer lies in the power of scale and the areas where generative AI technology can deliver the most value to the greatest number of users.

“There are an almost unlimited number of general use cases we can propose to our customers,” says Mueller. But with more than 300 million enterprise users using SAP cloud solutions, the German business software maker is looking for ways to horizontally extend AI to all types of customers.

One of the main areas of focus was SAP’s second generative AI pilot, Joule, which launched in September and by the end of this year. It will be fully integrated with all major SAP cloud applications, ranging from HR, Finance, Supply Chain and Procurement. Joule is already integrated with SAP’s HR SuccessFactors tool, helping customers who prepare 2 million job offers every year create better first drafts that can be used by hiring managers and recruiters.

Over the next two years, SAP will invest more than $1 billion in AI, focusing on products developed for customers – including workplace automation, data and analytics, and code development tools – while also embedding AI internally within SAP. In the first half of this year, SAP introduced more than 25 new AI capabilities that improve productivity and human resources tasks.

Last year, the company also invested in three generative AI companies: Aleph Alpha, Anthropic and Cohere. As the focus on artificial intelligence increases, SAP has reorganized, and the restructuring affects approximately 8,000 jobs that will either be filled voluntarily or reskilled internally.

SAP has taken an agnostic approach to partnerships, working with vendors from OpenAI to Anthropic and allowing developers using SAP products such as SuccessFactors to submit prompts to these large language models and compare the generated results. Mueller says that by experimenting with several models, usually at least two or three at a time, developers can find the model best suited to each task.

Mueller expects that more specialized large language models will be developed for more focused business applications. “I don’t think there will be hundreds of large language models that we will use in a business context,” Mueller says.

SAP develops its own models when it does not get what it needs from what is available in the market. Mueller says the company focuses specifically on tabular data, such as predictive analysis of financial data organized into rows and columns.

In April, SAP extended its contract with Mueller for another three years until the end of 2027, a common arrangement for the German company where the supervisory board has the power to advise executives like Mueller.

After the first two years of his contract, Mueller and the board meet to discuss his future path at SAP if he were to stay. “Planning is enough,” Mueller says. “I like it because it encourages good thinking and reflection.”

Mueller, who found himself on Fortune 40 2020 Under 40 List Joined SAP in 2013 and has held various leadership positions, including Chief Innovation Officer before becoming CTO in 2019. He oversees the SAP Business Technology Platform, which combines authoring and automation applications, data management, analytics and generative artificial intelligence capabilities into one unified environment for SAP applications.

Mueller believes that workplaces must allow employees to share ideas about how generative AI can automate tasks, rather than having management impose technology on the organization. To encourage more experimentation with generative AI, SAP created an AI playground last year that enabled more than 60,000 employees to try out more than 2 million AI suggestions.

“The best ideas often come from the people who do the work,” Mueller says.

John Kell

You can submit your thoughts and suggestions to CIO Intelligence here.

This story was originally published on Fortune.com