Organisations in Asia and Oceania are ahead of the global average when it comes to deeper AI adoption, but confidence in AI outputs, data availability and structured data remains uneven, according to Hitachi Vantara‘s latest State of Data Infrastructure research.

The finding lands as AI becomes embedded in more business-critical workflows, at a moment when many companies still lack confidence in the data available to support it. For Hitachi Vantara, that gap is the central story: deeper adoption does not necessarily mean businesses are ready to rely on AI across critical workflows.

Data foundations, not adoption, as the constraint

“AI Appreciation Day is a timely moment to recognise AI’s progress, but enterprise AI will not scale on adoption alone. With AI moving into more business-critical workflows, companies need to know that the data behind it is accurate, available and governed. Without that foundation, AI risks becoming another layer of complexity rather than a source of reliable business value. The next phase will be defined not by how much AI is being used, but by how reliably organisations can turn it into trusted business outcomes,” said Joe Ong, ASEAN Vice President and General Manager, Hitachi Vantara.

Ong’s point reflects a pattern increasingly echoed across the region’s enterprise technology sector: organisations have moved quickly to pilot and deploy generative and agentic AI tools, but the underlying data infrastructure – governance, quality controls, structured access – has not always kept pace with that enthusiasm.

A governance gap with regional implications

For Southeast Asian enterprises in particular, the gap carries practical weight. Regional regulators are paying closer attention to how AI systems are governed and audited, and unreliable data pipelines can undercut both compliance efforts and the business case for AI investment. Hitachi Vantara’s framing suggests that closing the application deficit, building the governance and data discipline needed to trust AI outputs, will matter more to enterprise AI’s next phase than simply expanding the number of use cases in production.