ODI and SAP partner to advance AI-ready data systems

The Open Data Institute and SAP have formed a partnership focused on AI-ready enterprise data infrastructure. The work will centre on research, governance and peer learning for organisations of different sizes.

The programme will examine how companies can prepare data for artificial intelligence systems, with a focus on trust, governance and interoperability. It will also bring together SAP customers, partners, policymakers and academics to develop research priorities and open standards.

The partnership comes as businesses try to apply AI tools to data systems originally built for transactions, compliance and reporting rather than machine use. That mismatch has created problems ranging from unreliable outputs to compliance risks when AI models are trained or deployed on poorly prepared information.

Under the arrangement, the ODI will help govern the programme through an independent model, drawing on its work in multi-stakeholder data initiatives. A second strand will produce research for Chief Information Officers and Chief Data Officers on approaches to making data AI-ready, including questions around machine learning, generative AI and agentic AI.

A third area will focus on building a community around the programme. The aim is to create a forum where businesses, policymakers and researchers can share practice and contribute to common frameworks for enterprise data.

Research focus

The research agenda will examine several data management models already in use, including data lakes, data mesh, data fabric and data products. The programme is intended to explore how those approaches interact with newer AI systems and what governance structures are needed to support them.

A research programme board will oversee the project, steer priorities and publish findings. The partners also plan to gather industry and academic input as the work develops.

Louise Burke, Chief Executive Officer of the ODI, outlined the case for the partnership in a statement.

"AI will define enterprise competitiveness for the next decade, but competitive advantage doesn't come from AI models alone. It comes from the quality, governance, and autonomy of the data beneath them. Most organisations are sitting on data that simply isn't ready for AI, and the consequences of getting this wrong, from biased outputs to regulatory non-compliance, are significant. Through this partnership, the ODI and SAP are bringing together the expertise, research, and community needed to give organisations the blueprint they need. Our goal is to make AI-ready data infrastructure accessible to enterprises of all shapes and sizes, built on open standards that no single vendor controls," said Louise Burke, Chief Executive Officer, ODI.

Her comments point to a wider issue for large organisations: while attention has often centred on AI models and applications, practical deployment depends on the quality, lineage and governance of the underlying data. For many companies, that means revisiting older data estates and deciding how to adapt them for newer AI uses.

Data trust

SAP framed the issue in similar terms, arguing that the main barrier to scaling AI is not access to models but confidence in the data those systems use. Organisations with governed and integrated data are better placed to achieve measurable business results from AI deployments, it said.

"As companies scale AI in 2026, the real gap is data trust, not technology - organisations with governed, integrated data are far more likely to outperform and deliver measurable results. A critical next step for companies is to establish a business data fabric, which can ensure AI agents have the context they required to understand the business and impact," said Irfan Khan, SAP Chief Product Officer for Data & Analytics, SAP.

The ODI is a London-headquartered non-profit founded by Sir Tim Berners-Lee and Sir Nigel Shadbolt. It works with businesses, governments and public bodies on data practice, research and policy, with a focus on open and trustworthy data ecosystems.

The partners are also seeking wider sector involvement and funding for the project as they develop what they describe as an open and interoperable blueprint for enterprise data to support AI use.