SAP Sapphire 2026 unveils agentic AI platform
SAP Sapphire 2026, held in Orlando, Florida, delivered a torrent of product and service announcements centered on what the company calls the “new SAP.” The event showcased a comprehensive agentic AI infrastructure ecosystem, with executives highlighting tools that enable secure, compliant, and scalable enterprise AI deployment. From CTO Philipp Herzig to BTP president Michael Ameling and other practice leaders, the message was clear: SAP is building the backbone for an autonomous enterprise.
Addressing the enterprise AI gap
Philipp Herzig, SAP’s chief technology officer, opened discussions by addressing the so-called “enterprise AI gap.” He noted that while tools like Claude Code offer phenomenal capabilities, customers struggle to make live AI secure, compliant, and manageable over its lifecycle. “The problem is that customers have trouble making live deployed AI secure, compliant and capable of working within the realms of its own lifecycle management,” Herzig said. This challenge is magnified at scale, where systems risk brittleness. Herzig emphasized that SAP’s platform provides the backbone for business data context, allowing teams to build tools with the right processes to achieve reliable outcomes.
The enterprise AI gap, explained Herzig, stems from the difficulty of moving from experimentation to production. Many organizations find that out-of-the-box AI solutions are prone to hallucinations or lack governance. SAP’s approach spans multiple tiers of the technology stack, ensuring that agents and workflows run securely across business functions from finance to procurement.
Three-layer architecture for agentic AI
SAP’s new architecture consists of three distinct layers. First, the process layer, which spans applications within an integrated suite where agents and workflows execute across every business domain. Second, the product-level data and AI fabric, which provides the entities and business objects underpinning processes. Third, the operational layer model, a semantic domain that defines where teams work. This structured approach enables developers to build agentic services with confidence, knowing that all plumbing and lower-level engineering is handled. “Because so many developers may not even know what they should be building to improve business processes and KPIs, we also make sure SAP Signavia is there to help provide a process consulting agent,” Herzig added.
SAP BTP and agentic function democracy
Michael Ameling, president of SAP Business Technology Platform (BTP), discussed the company’s vision for democratizing agentic functions. He stated that SAP wants to remove barriers to using agentic services across the ecosystem. “That’s why everything has been intelligently wired together so that software application developers and businesspeople alike can both interact with services to drive the autonomous enterprise forward,” Ameling said. He highlighted that agentic controls now enable users to “just log on” while infrastructure oversees identities and policy access. IT remains in control with observability metrics aligned to the OpenTelemetry (OTEL) standard. Notably, SAP is the ninth largest contributor to open source code commits globally, reflecting its commitment to community standards.
Efficient AI deployment: core workflows vs. advanced analytics
Ameling further emphasized the importance of efficiency in AI deployment. He argued that for deterministic processes like invoice processing, core algorithmic logic should be used rather than burning tokens on large language models. “If a company needs to drive invoice processing functions, that’s essentially a deterministic process with a start, middle and finish,” he said. Only when non-deterministic nuances arise—such as assessing whether invoices align with changing company policy—should AI be brought in. This approach balances cost and performance, enabling a more responsive enterprise.
Hands-on demos: SAP-RPT-1 and agentic workloads
Richard Grandpierre, vice president of product management for SAP Business AI, provided practical demonstrations of the new capabilities. He traced the evolution from predictive and prescriptive AI to today’s agentic workloads, noting that while the pace of change is rapid, agentic AI is here to stay. “Agentic workloads actually do work—they perform tasks, they use tools, they manipulate data and so much more,” Grandpierre said. He showcased SAP-RPT-1, a tabular foundation model designed for structured business data. Unlike language models, RPT-1 works with complete tables and fields within ERP environments to autocomplete and suggest values for real-time business operations. This reflects SAP’s early recognition that the greatest enterprise AI opportunity lies not in LLMs but in AI built for structured data.
Interoperable AI agents across ecosystems
Balaji Balasubramanian, CX Lead at SAP, focused on customer experience and the partnership with Google Cloud. He argued that SAP has transitioned from systems of record and transaction through systems of engagement to now systems of agentic engagement. Integrations between SAP Engagement Cloud, SAP CX, Joule, and Google’s Gemini Enterprise allow marketers to deploy agents that securely access unified data across ecosystems. “This is more than a data integration; it’s a leap forward for AI agents that can collaborate naturally and execute seamlessly,” Balasubramanian said. The partnership aims to move organizations from AI experimentation to AI-enabled customer experience at scale, reducing manual tasks and freeing marketers to shape customer journeys.
Joule: the voice of the autonomous enterprise
The event’s main keynote closed with Joule, SAP’s AI voice, summarizing the theme: turning the promise of business AI into reality. The SAP Business AI platform provides the data, process, and governance needed for accurate, secure outcomes at scale. SAP introduced the Autonomous Suite, where applications reason, decide, and act on behalf of users. Together with customers and partners, the company demonstrated how Rise with SAP and change management tools enable the vision of the autonomous enterprise. Joule’s closing remarks underscored the transformative potential of agentic AI when built on a trusted, integrated platform.
Source: Computerweekly News