SAP is putting AI directly under its top technology leadership, but you should judge the reshuffle by what customers can buy, measure, and renew, not by how clean the new org chart looks.
SAP has a simple problem now. It has told investors and customers that Business AI is central to its future, so the people running the company have to look as if they can deliver it. In early March, Bloomberg reported that CEO Christian Klein reorganized SAP's executive board around an explicit "all in on AI" push, with Philipp Herzig moving from Chief AI Officer to Global Chief Technology Officer and Thomas Saueressig becoming Chief Customer Officer over a combined Customer Value Group.
That is not a small reshuffle at Europe's most valuable software company. Herzig now carries the AI strategy and technology brief together. Saueressig gets the customer-facing machinery: success, services, deployment, and the harder work of proving value after the contract is signed. Muhammad Alam, who had been responsible for product engineering, said he won't renew his contract when it expires in March 2027, citing personal reasons.
Look closely at that last point. Alam is the executive most directly tied to the product engineering base under SAP's Business AI promise. If you're a customer running finance, procurement, supply chain, or HR processes on SAP, you don't really care whether the board structure sounds modern. You care whether Joule answers something useful, whether the AI sits inside the workflow you already use, and whether the price is clear before renewal season starts.
SAP's share price has already had a rough lesson in investor patience. The company's Q4 2025 cloud backlog grew 25%, but came in below guidance and triggered what analysts described as SAP's worst single-day drop since 2020. The stock has recovered some ground since then, but HSBC flagged the issue SAP can't talk its way around: roughly 60% of its on-premises customer base still hasn't started moving to the cloud. That is a huge pool of future revenue. It is also a huge pool of customers with every reason to wait.
According to SAP's own SEC filings, cloud revenue grew 27% at constant currencies in Q1 2026, and Business AI was embedded in two-thirds of Q4 cloud order entries. Those numbers give Klein something real to point to. They also raise the bar. Once AI is part of the sales motion, customers will start asking what it actually saved, accelerated, or automated. A board title won't answer that.
Microsoft, Oracle, and Salesforce are not politely standing aside while SAP explains its transition. Microsoft Copilot already sits inside the office tools most SAP customers use every day. Oracle has pushed AI across Fusion Cloud and has been aggressive around renewals. Salesforce's Agentforce is aimed straight at customer-facing workflows, the same terrain Saueressig's new Customer Value Group is supposed to defend.
SAP's strongest answer is also its oldest one: it owns the transactional system of record in some of the most important business processes on earth. No hyperscaler has SAP's depth in purchase orders, invoices, warehouse flows, and month-end finance work. Frankly, Klein is right to lean into that. But being right about the data is not the same thing as shipping useful product.
The product test is coming fast
At SAP Sapphire on May 13, 2026, Klein presented the company's "Autonomous Enterprise" vision, and TechRadar reported that SAP introduced the SAP Business AI Platform, SAP Knowledge Graph, Joule Studio, Joule Work, and the SAP Autonomous Suite. The names are a mouthful, but the aim is clear enough: give AI agents enough governed business context to act across SAP systems without turning mission-critical operations into a guessing exercise.
That phrase, mission-critical, is doing real work here. If an AI tool drafts a bad sales email, you fix the sentence. If it mishandles a procurement approval, finance reconciliation, or supply-chain exception, you may have a far more expensive problem. Klein's point at Sapphire was that "almost right" isn't good enough for enterprise operations. Customers will agree. They will also expect SAP to prove it.
The pricing change makes the pressure sharper. Bloomberg reported in March that Klein said SAP would begin charging for AI consumption through "AI Units" rather than relying only on per-user subscriptions. For SAP, that could tie revenue more closely to usage. For customers, it introduces a new billable meter into software budgets that are already under scrutiny. If you run procurement for a large manufacturer, you now have to ask a direct question: what happens to annual spend if AI agents start doing more work inside the system?
That is where Saueressig's new role starts to make sense. SAP is not only selling cloud migration now. It is selling a measurable AI outcome layered onto cloud migration, and those are different promises. A customer success team can smooth deployment. It cannot manufacture value if the product doesn't produce it.
The Herzig promotion sends a clearer signal. SAP no longer wants AI treated as a feature set bolted onto ERP. It wants AI and core technology strategy owned together. That is the right structure if SAP can keep product execution disciplined after Alam's planned exit. CIO.com noted that SAP's new Extended Board is meant to help with coordination below the executive board, but an advisory layer is not the same as a single product owner with full accountability.
SAP has roughly 30,000 enterprise customers globally, many of them already stuck in long, expensive migration programs from on-premises systems to S/4HANA Cloud. They don't need another declaration that AI is strategic. They need the invoice dispute resolved faster, the supply-chain exception handled correctly, the close process shortened by a measurable number of days.
Klein told Fortune in May that the AI race "is being fought in the wrong place," arguing that operational context matters more than model capability. He is not wrong. The catch is brutal and simple: SAP has to turn that context into working software before its customers decide the AI surcharge is just another line item.
Also read: Indonesia sentences Gojek co-founder Nadiem Makarim to 10 years in prison over Chromebook procurement corruption • Chamath Palihapitiya steps back into the operator seat as 8090 Labs closes a $135 million Series A to chase the enterprise AI coding market • Ferrari and BMW are ditching copper wiring, and the implications run well beyond the factory floor
