Copyright Business Insider
To illustrate the stakes of data integrity and accuracy in enterprise AI, Manos Raptopoulos of SAP likes to pose a simple question: "Have you ever tried to do a word count using a generative AI tool?" Compared to the text's true word count, such a tool today might be 10% off. After improvement, that might come down to 5%. But then impose that degree of imprecision on your EBITDA, said Raptopoulos, who oversees Asia, Europe, and Middle East Africa for SAP as chief revenue officer and is a member of the company's extended board. "In the business world, a difference of that magnitude can influence analyst expectations, it can influence valuations," he said. "It's a nightmare for the CFO not to have a deterministic way of saying, 'this is my number.'" The journey from probabilistic to deterministic With AI seamlessly embedded into SAP's applications and fueled by the world's most powerful business data, the SAP Business Suite is unparalleled in its ability to deeply connect every part of a customer's business, enabling companies to achieve real business outcomes and end-to-end transformational value. AI-powered enterprise apps need to balance the deterministic capabilities of machine learning with the more probabilistic methods of generative AI, Raptopoulos said. That's why he advocates for enterprise access to a data pool that's as diverse and governed as possible. SAP customers represent 84% of the total global economy. With a footprint in more than 180 countries, SAP brings deep knowledge of compliance and regulations across regions. This collective expertise and insight ladder up into an unrivaled intersection of applications, data, and AI, each driving the other forward, Raptopoulos said, noting this convergence positions the SAP Business Suite as a game-changer. SAP's approach to data anticipates concerns shared by IT professionals. In the CIO Vision 2025 report — a survey of 600 CIOs across 18 countries and 14 industries — Databricks and MIT found that 72% of respondents consider their data-quality protocols "reactive" or merely "aware," highlighting a tendency to view data management as a tech partner's responsibility rather than an internal concern. AI's performance and reliability are only as good as its data — and SAP's depth of data is unmatched. Raptopoulos feels a mandate to help organizations create tangible business outcomes by leveraging the SAP Business Data Cloud, which harmonizes data from SAP and non-SAP systems. The integrated data is then transformed into a centralized, semantically enriched layer to fuel advanced analytics and AI capabilities. "That last mile," he said, "is super important to get right." The balancing act between data and productivity Being a few degrees off target is also a condition the aerialists of SAP customer Cirque du Soleil Entertainment Group seek to avoid. Manual recordkeeping and travel management for 7,000 trips per year was a particular pain point for the global franchise. Cirque du Soleil finance managers needed to manually create and track around 140 cost objects for each tour across a slate of 50 discrete shows. The SAP Business Suite cut through this complexity, creating a unified data model and intelligent technology to negotiate better rates with suppliers. Finance teams reduced the amount of cost objects created per tour by nearly 80%. Cirque du Soleil also uses SAP tools to improve data integrity and governance, and to simplify materials management and production planning for the sweeping collections of costumes the company designs and manufactures each year. Through its reliance on SAP, "our workload is cut in half," said Chantale Périgny, accounts payable manager for Cirque du Soleil. "Data flows easily and transparently through the system, so we can focus more on control than the process itself." Kickstarting a "virtuous cycle" For Raptopoulos, this example illustrates what customers should expect from SAP: The ability to overcome complexity, innovate faster, and achieve sustained success in a rapidly evolving world by integrating applications, data, and AI into one unified system. Circumventing a black-box model where users don't know how models were trained, SAP's Business Data Cloud creates what Raptopoulos described as a "virtuous cycle": a foundational model of quality data, expanded upon when customers trust and use it, bringing even more data into the overall ecosystem. This cycle fosters innovation, Raptopoulos said, because users feel empowered to experiment with new AI applications to solve pain points by more easily surfacing insights from vast datasets. Joule, SAP's AI copilot, adds value by adding context. When Joule uses its natural-language processing ability to address a customer question, "it's quite important to understand, is this a finance question? Is this a supply chain question?" Raptopoulos said. "The nuances can be quite different and that's where a well-trained gen-AI model should instantly understand the context." Those interactions accrue into insight that saves time and money, Raptopoulos said — gains that create momentum for growth. "If the context is there, the semantics are there, your business can access the outcomes AI promises to deliver. That's the virtuous cycle, because you then use that to train more models and create more meaningful applications."
