SAP CFO says AI must move beyond chatbots for real returns

SAP’s finance chief said artificial intelligence in enterprise software has to move beyond chatbots and coding assistants before companies can expect real payoffs from their spending. Dominik Asam said the “lion’s share” of AI token consumption today is still going to “low-hanging fruit,” a blunt warning after SAP’s second-quarter results.

Asam’s point lands at a moment when executives are under pressure to show that AI spending is producing more than faster answers and tidier support desks. He said the harder task is embedding AI into more complex business processes, where clean, governed data can be used inside core workflows rather than treated as a standalone conversational layer. That means using AI in planning, coding, customer service, procurement, analytics and decision-making, not just as a chatbot interface.

The comments also sharpen a broader question hanging over the corporate AI cycle: when do heavy investments in infrastructure, software and talent start turning into measurable productivity gains, higher margins or new revenue streams? SAP’s own research page says AI could drive a 31% return on investment and estimates the average global business is spending US$26.7 million on AI this year, figures that underscore both the scale of the bet and the pressure to prove that returns are more than incremental.

SAP’s position carries unusual weight because its software sits in the middle of mission-critical operations across finance, procurement, supply chain, human resources and customer experience. That makes the company a useful test case for where AI can move beyond novelty. SAP’s public profile says it helps organizations run those functions, and that is exactly where Asam wants customers to push AI deeper, into systems that touch real transactions and decisions rather than isolated prompts.

The distinction matters because basic chatbot deployments can shave time, but they do not automatically alter the economics of a business. Meaningful returns for large employers would look different: fewer manual handoffs in procurement, faster close cycles in finance, better demand planning, lower support costs, and analytics that improve decisions without adding risk. Asam’s message was that the easy wins may already be priced in, and the next phase will be judged on whether AI can work inside the machinery of the enterprise.