SAP and the Open Data Institute (ODI) have launched a global program to help organizations design “AI‑ready” data foundations, signaling that the hardest work in enterprise AI is now squarely in architecture, governance, and interoperability rather than in model selection alone.
The Open Data Institute (ODI) is an independent, non-profit organization co-founded in 2012 by Sir Tim Berners-Lee and Sir Nigel Shadbolt to promote the use of open data to build a trustworthy, efficient data ecosystem.
Sir Tim Berners-Lee is an English computer scientist best known as the inventor of the World Wide Web, HTML, the URL system, and HTTP. Prof. Sir Nigel Shadbolt is a leading researcher in artificial intelligence and was one of the originators of the interdisciplinary field of web science.
Headquartered in London, U.K., the ODI provides training, research, and consultancy to help governments and businesses maximize data benefits.
The collaboration underpins IDEA (Interchange for Data and Enterprise AI), a multi‑stakeholder effort to define what good looks like when AI systems span multiple platforms, vendors, and regulatory regimes.
From ERP Data Exhaust to AI‑Ready Infrastructure
Most large enterprises already sit on decades of ERP and line‑of‑business data, but that estate was never designed with AI workloads in mind. Systems like SAP S/4HANA, Oracle Cloud ERP, Microsoft Dynamics, and legacy on‑premises estates have been optimized for transactional integrity, compliance reporting, and human decision‑making, not for large‑scale model training or agentic AI that traverses process silos.
The SAP–ODI partnership acknowledges that gap explicitly. Rather than pitching another “AI layer,” it focuses on the less glamorous work: shared semantics, policy‑driven access, lineage, and data products that can be trusted by AI systems regardless of which ERP or data platform they originate from.
Louise Burke, CEO of ODI, said, “AI will strongly influence the competitiveness of organizations over the next ten years. However, the difference lies not only in the models, but precisely in the quality, governance, and autonomy of the underlying data. Many organizations possess data that is not yet suitable for AI. If this backfires, it can lead to distorted outcomes and regulatory issues.”
That framing matters in a market where most enterprises now run mixed estates – SAP at the core of finance and supply chain, perhaps, but surrounded by other ERPs, best‑of‑breed SaaS, data platforms, and industry clouds.
Analysis
What This Means for ERP Insiders
Clean core meets messy reality. Many S/4HANA programmes and other ERP modernisations are justified in terms of “clean core” and composable architecture. An AI‑ready data agenda forces programme teams to decide what actually belongs in the ERP, what becomes a data product at the edge, and how those elements are governed when AI agents start orchestrating processes across systems.
IDEA: A Neutral Interchange, Not a New Platform
IDEA is positioned as an interchange rather than a new product: a governance and research framework that brings together vendors, customers, policymakers, and academics to define and test patterns for AI‑ready data. ODI’s role here is to provide a neutral backbone for the work – drawing on its history convening multi‑party data initiatives – while SAP contributes its experience with complex enterprise data models and “business data fabric” concepts.
Daniel Dukes, senior director of product marketing at SAP, wrote in a blog post, “A business data fabric brings heterogeneous sources into a consistent state, making it easier to combine SAP and non-SAP data. For data scientists and AI engineers, efficient data access and trusted AI can become the rule, not the exception. Ultimately, operationalizing a data fabric shortens the project cycle—delivering faster time to value,”
For ERP leaders, that matters in three ways:
- It acknowledges multi‑vendor reality. Even though this is a partnership with SAP, most of the AI‑relevant data in large organizations still straddles multiple ERP systems and data platforms. A neutral program offers a forum to confront that complexity.
- It treats governance as design, not afterthought. The focus on independent governance signals that risk, accountability, and explainability are core design inputs for AI‑ready data, not compliance wrappers added at the end of a project.
- It creates space for patterns by looking explicitly at how classical ML, generative AI, and emerging agentic approaches stress different data architectures (lakes, mesh, fabric, domain products). IDEA has a chance to produce reference models that CIOs and CDOs can actually implement.
Analysis
What This Means for ERP Insiders
Data architecture becomes the battleground. ERP buyers have heard for years that “data is the new oil,” but the real contest now is over who gets to define the reference architectures and standards that shape how that data is exposed to AI. If SAP and ODI can anchor credible patterns for AI‑ready data, it will influence how other vendors – from hyperscalers to rival ERP and data‑platform providers – position their own frameworks
What to Watch Next
For now, the partnership is at the governance and research stage, but it sets up several concrete developments ERP leaders should monitor.
One key area to watch is how quickly the program moves from abstract principles to reference architectures, patterns, and templates that can be implemented across major ERP and data platforms. The speed of this transition will determine how actionable the initiative becomes for enterprise teams.
Another critical factor is whether other large vendors, especially hyperscalers, rival ERP providers, and data-platform specialists, participate meaningfully or respond with competing frameworks. Their involvement will shape whether IDEA becomes a unifying standard or one of several parallel approaches.
Finally, attention will turn to how regulators and industry bodies engage with the outputs. If IDEA’s recommendations begin to appear in sectoral guidelines, audit frameworks, or procurement criteria for AI-enabled systems, it would signal broader institutional adoption and long-term impact.
For organizations trying to turn ERP‑adjacent data into reliable AI outcomes, the message is clear: AI readiness starts in the data and governance layers, not in the model catalogue. This SAP–ODI program won’t solve that for you, but it may help define the playbook everyone will be judged against.
Analysis
What This Means for ERP Insiders
Regulation and trust will shape ERP roadmaps. As AI regulation in Europe and elsewhere hardens, boards will ask awkward questions about where training data came from, how models behave for different customer segments, and how quickly problematic outputs can be traced back and corrected. Programmes like IDEA could influence the checklists that regulators, auditors, and industry bodies expect ERP‑centric organisations to meet.
