At SAP Sapphire 2026 this year, one thing became obvious: SAP is fully committing to AI as the future of the enterprise. From the expansion of Joule to the announcement of the Autonomous Enterprise, Sapphire 2026 focused heavily on how AI agents, automation, and semantic business context will reshape the way organizations operate. But there was a more important reality emerging: most enterprise data still is not ready for AI. And that may become the biggest barrier to enterprise AI adoption over the next few years. Semantik, the new platform from EPI-USE Labs, is a trusted semantic foundation to help organizations prepare data for practical, governed, reliable enterprise AI
SUMMARY: SAP Sapphire 2026 showed SAP’s shift from AI Assistants to AI Execution, with Joule, the AI Agent Hub, Knowledge Graph, Company Memory, and Domain Models supporting the Autonomous Enterprise. But enterprise AI has a data problem: most SAP environments are not AI-ready because they lack semantic business context. Semantik, the new platform from EPI-USE Labs, is a trusted semantic foundation to help organizations prepare data for practical, governed, reliable enterprise AI.
At SAP Sapphire 2026 this year, one thing became obvious: SAP is fully committing to AI as the future of the enterprise.
From the expansion of Joule to the announcement of the Autonomous Enterprise, Sapphire 2026 focused heavily on how AI agents, automation, and semantic business context will reshape the way organizations operate.
But underneath all the keynotes and product announcements, there was a more important reality emerging across customer conversations: most enterprise data still is not ready for AI.
And that may become the biggest barrier to enterprise AI adoption over the next several years.
The shift from AI Assistants to AI Execution
This year’s Sapphire messaging marked a major evolution in SAP’s AI strategy.
The focus is no longer just on AI copilots helping users search for information or summarize reports. SAP is now positioning AI as an operational layer capable of coordinating and executing business processes across the enterprise. They've given the architecture a name: the SAP Business AI Platform – Joule, the AI Agent Hub, Knowledge Graph, Company Memory, and Domain Models, under one governance fabric.
SAP calls this the ‘Autonomous Enterprise’.
The vision is ambitious:
- AI agents coordinating workflows across finance, procurement, HR, and supply chain
- Intent-driven interactions through Joule
- Industry-specific AI models
- Agent-led transformation and automation
- Enterprise systems capable of acting, not just responding
It’s a significant shift in enterprise software strategy.
But for AI to operate effectively at scale, it needs far more than model power alone. It needs business context.
Enterprise AI has a data problem
One of the biggest misconceptions in enterprise AI is that moving data into the cloud automatically makes it AI-ready.
It doesn’t.
Most SAP environments were never designed for AI consumption. Years of duplicate records, bloated custom Z tables, fragmented storage, disconnected metadata, and commingled sensitive information make it extremely difficult for AI systems to interpret enterprise data accurately.
That creates what we see as the enterprise AI paradox:
AI requires rich, contextual, high-fidelity data to generate meaningful outcomes, but most ERP environments lack the semantic structure needed to provide that context.
Without it:
- AI models hallucinate or infer incorrect relationships
- Business meaning gets lost across disconnected tables
- Data quality issues multiply
- Governance and privacy risks increase
- Operational costs continue growing alongside bloated environments
This theme showed up repeatedly throughout Sapphire – in SAP's own Knowledge Graph and Company Memory announcements, in the AI governance and trusted-data messaging, and across client conversations.
AI needs business context, not just data
As organizations move from AI experimentation toward operational deployment, the focus is increasingly shifting toward the foundational work required to make enterprise AI practical, governed, and reliable within real SAP environments.
Our Head of Product Engineering, Dr Tiaan Scheepers, describes it as an iceberg. Above the waterline: SAP's Business AI Platform – Joule, agents, the visible AI everyone talks about. Below the waterline: your enterprise data – decades of SAP semantics, custom Z-tables, payroll cycles, hierarchies – the substrate the platform actually has to reason on. You need a solution to ensure that what's below is ready for what's above.
That’s also why semantic mapping, structured business object relationships, secure non-production data, and clean core strategies are becoming increasingly important as organizations prepare SAP environments for AI.
Introducing Semantik by EPI-USE Labs
We brought Semantik to Sapphire in person. At our annual Oceanaire reception during Sapphire week, nearly 200 clients and prospects joined us with real conversations about what AI-ready data actually means. Semantik, our AI-native data platform, creates a trusted semantic foundation across SAP and broader enterprise landscapes. The timing wasn't an accident: the gap SAP described from the keynote stage is the exact gap Semantik was built to close.
Semantik was built around a simple but increasingly urgent challenge: Enterprise AI cannot succeed if enterprise data lacks semantic meaning.
Most enterprise AI projects fail. Not because the models are bad – because the data they're reasoning on has no meaning.
Paul Snyman
Head of Product Management, EPI-USE Labs
Unlike many AI platforms starting from scratch, Semantik builds on more than 42 years of SAP expertise and the trusted data management foundations already used by over 1,900 organizations globally. Using thousands of predefined SAP business object definitions and semantic relationships, the platform helps connect and structure fragmented enterprise data into a business language AI systems can actually understand.
According to Dr Tiaan Scheepers, Semantik represents the semantic layer EPI-USE Labs has been building inside SAP customer deployments for more than 20 years, now packaged as a foundation for enterprise AI and agentic automation.
That semantic layer allows organizations to move beyond technical tables and codes, enabling more natural interaction with enterprise systems, automated workflows, and more accurate AI-driven decision making.
Semantik is designed to integrate natively with SAP Joule so capabilities like test data provisioning, data privacy processes, and data masking can be triggered directly within AI-driven workflows as those integrations roll out.
The timing of the launch reflected many of the themes discussed throughout Sapphire, particularly around AI readiness and semantic business context.
EPI-USE Labs' colleagues celebrate the launch of Semantik at SAP Sapphire in Orlando
The biggest takeaway from Sapphire 2026
The AI vision presented at Sapphire was compelling. But the organizations that will benefit most from enterprise AI won’t necessarily be the ones adopting AI tools the fastest.
They’ll be the ones that:
- clean up and structure their data
- reduce unnecessary complexity
- establish semantic business context
- modernize intelligently
- govern AI responsibly
- create trusted foundations for automation
Because AI is only as powerful as the enterprise understanding behind it.
The organizations that solve the data and semantic context problem first will be the ones that get real value from enterprise AI.