SAP CFO: Enterprise AI Must Move Beyond Chatbots to Deliver Real Returns

SAP Chief Financial Officer Dominik Asam issued a direct challenge to enterprise AI leaders on July 23: stop treating AI as a chatbot layer and start embedding it in core business workflows. Speaking after SAP's second-quarter earnings, Asam said corporate investments in generative AI have not yet yielded broad productivity gains, and the reason is where companies are deploying it.

Most AI token consumption, Asam noted, remains focused on lower-risk tasks such as coding assistance and basic chat interfaces. The real value, he argued, comes from AI embedded within specific, governed business processes, not from general-purpose models bolted on top of existing systems.

Key Takeaways

• Sequential errors compound in core workflows: Asam specifically flagged finance as a domain where AI hallucinations carry outsized risk. In a multi-step financial process, errors compound statistically. "It requires much more excruciating assurance levels," he said.
• Data quality is a prerequisite, not an afterthought: Asam was direct that AI will not fix legacy data silos. "The idea that AI will solve all these problems if they are messy, legacy data silos is not true," he said, adding that poor data architecture also drives "extremely high token costs."
• Enterprises will choose on cost and reliability: Asam said businesses will ultimately favor the most cost-effective and reliable tool, whether that means standard software, open-source applications, or high-cost frontier models. Vendor lock-in to premium models is not a given.

The signal for enterprise AI practitioners: deployment in core functions requires cleaning and governing internal data first, and the governance standards for AI in finance or supply chains are meaningfully higher than those for productivity tools and chatbots.

Read the full article on Sekber News