AI adoption across Asia-Pacific is accelerating at a pace that few industries have experienced before. According to Denodo’s 2025 Market Study on Modern Data Architecture in the AI Era, more than 80% of enterprises plan to deploy modern data architecture by the end of 2025. The reason is simple: AI is only as powerful as the data that fuels it. Enterprises now recognise that the way data is structured, governed, and accessed determines how effectively AI can deliver real business value.
This shift marks a fundamental rethinking of how organisations view data strategy. What was once an IT issue has become a board-level priority that defines agility, risk management, and competitiveness. As regulations evolve and AI workloads expand, the ability to unify data across complex, distributed systems is no longer optional. For many organisations, the focus has shifted from rebuilding infrastructure to modernising intelligently, using flexible, outcome-driven architectures that scale without disruption.
“CIOs and CFOs alike are now demanding measurable outcomes within months, not years, and that is fundamentally changing how data architecture is planned,” said Shanmuga Sunthar Muniandy, Denodo’s Director of Data Architecture and Chief Technology Evangelist for APAC. The pressure for faster results and accountability is driving a region-wide transformation in how enterprises approach both technology and investment.
From large-scale overhauls to outcome-first, composable stacks
The traditional model of large, multi-year data transformation projects is being replaced by incremental, composable approaches that deliver value more quickly and with less risk. Shanmuga Sunthar, who has worked in data management and analytics for over three decades, explained that these smaller initiatives have consistently proven more successful. “Success comes when one starts with small yet strategic data initiatives and scale incrementally from there,” he said.
The logic is clear. Rather than commit to long-term overhauls that can quickly become outdated, enterprises are focusing on smaller, outcome-driven projects that align with evolving business needs. Denodo’s research supports this trend. Nearly half of the surveyed organisations now spend less than US$100,000 on their data architecture initiatives, compared with just 11.4% in 2023. This indicates a decisive move away from capital-intensive rebuilds towards modular investments that can deliver early ROI and adjust as technologies mature.
Composable architectures make this possible. Instead of replacing legacy systems entirely, enterprises are integrating modern components such as lakehouses, data fabrics, and semantic layers to build flexible stacks that meet specific goals. Logical data management serves as the connective tissue between these systems, providing a unified, governed layer of access without requiring data duplication. This approach preserves existing investments, accelerates delivery, and ensures that businesses remain agile as they scale.
Outcome-driven design also changes how success is measured. Instead of focusing on completion milestones, enterprises can track performance against tangible outcomes such as faster insights, greater accuracy, or improved compliance. This builds confidence across executive teams, demonstrating that modern data architecture can deliver measurable business impact without disrupting operations.
Solving the last mile: Trusted, governed data for AI adoption
Even as data infrastructure improves, many AI projects struggle to move from the pilot phase to full deployment. The reasons are often less about technology and more about governance, trust, and accessibility. “Lineage, explainability, and governance remain major hurdles in AI operationalisation,” said Shanmuga Sunthar. Without the ability to trace and validate the data used to train AI models, organisations risk compliance breaches, poor performance, and loss of stakeholder confidence.
Logical data management addresses these challenges by creating a virtualised access layer that connects diverse data sources in real time. This unified layer ensures data quality, semantic consistency, and policy enforcement across systems. When combined with lakehouses, it allows enterprises to centralise metadata and control data governance through a single, coherent framework. The result is greater transparency and reliability across AI use cases.
The Denodo and Lakehouse ROI Report highlights that enterprises adopting this hybrid approach achieve significant operational savings, including a 77% cost reduction in maintaining governance and consistency. Beyond the financial benefits, this architecture provides explainable and auditable data that supports responsible AI practices. This capability is increasingly essential in the Asia-Pacific region, where governments are tightening privacy laws and introducing AI accountability frameworks to protect consumers and businesses alike.
Self-service access to governed data is another key benefit. Logical data management allows business users and analysts to explore and act on reliable data without depending entirely on IT or engineering teams. This speeds up model development, shortens experimentation cycles, and enables more agile innovation across departments. With these capabilities, enterprises can move AI from proof of concept to production at scale, ensuring systems are both compliant and operationally effective.
Proving ROI beyond cost control
As modernisation efforts mature, leaders are re-evaluating how to measure success. ROI is no longer limited to cost savings or infrastructure efficiency. The focus has expanded to include agility, innovation, and the ability to turn data into measurable business outcomes. Shanmuga Sunthar noted that assessing ROI for modern data initiatives must move beyond traditional metrics to capture enterprise agility, operational efficiency, and strategic value.
Independent studies reinforce this shift. Research conducted by Veqtor8 found that Denodo-enabled architectures deliver 345% ROI over 3 years and achieve payback in under 7 months. Engineering effort is reduced by 75% and implementation costs by 78% through reusable connectors and semantic abstraction. Most importantly, time-to-insight improves by up to four times, enabling faster decision-making and greater responsiveness to market changes.
Strong governance and compliance also contribute to ROI by reducing risk exposure. Enterprises that can demonstrate control and transparency over their data assets gain both regulatory assurance and customer trust. These factors translate into measurable business value, supporting sustainable growth and improved reputation.
Modern measurement frameworks are now multi-dimensional. They consider financial performance alongside operational efficiency, innovation capability, and risk mitigation. This allows CIOs and CFOs to justify investments not as technology expenses but as strategic enablers. By linking data agility directly to business agility, organisations are creating a new model for evaluating digital transformation—one that aligns technology outcomes with overall enterprise value.
This mindset is also reshaping how data teams are perceived. Once seen primarily as back-end support, they are now central to driving measurable performance improvements. The ability to connect data initiatives to commercial results, from new product development to improved customer retention, is redefining the role of data leadership across industries.
Future-proofing under multi-cloud and regulatory complexity
Enterprises in Asia-Pacific face an increasingly complex business environment defined by diverse regulatory systems, rapid technology shifts, expanding multi-cloud ecosystems, and evolving customer expectations. Designing a data architecture that can absorb these changes has become a strategic necessity.
Shanmuga Sunthar described the region as one of “diverse regulatory, technological, and business volatility,” where flexibility is critical for long-term success. The solution lies in modular, policy-driven design supported by logical data management. This enables centralised governance and access control, ensuring that the right people can access the right data under the right conditions. It also allows privacy, residency, and usage rules to be applied dynamically across multiple environments without data duplication.
Policy-as-code and metadata-driven governance frameworks further enhance adaptability. By embedding rules directly into data workflows, organisations can achieve continuous compliance and automated enforcement. This makes it easier to respond to evolving standards such as Singapore’s PDPA, Australia’s Privacy Act, or new AI accountability laws emerging across the region.
At the same time, enterprises are increasingly focused on avoiding vendor lock-in. Open file formats, standard query interfaces, and containerisation technologies ensure interoperability and portability across platforms. This flexibility allows teams to experiment with emerging AI tools or migrate workloads without disrupting existing systems.
“Future-proof data platforms are not about anticipating every rule; they are architected to absorb change,” said Shanmuga Sunthar. “Continuous auditability enables organisations to respond quickly to regulatory or AI accountability inquiries.” This philosophy underpins a new generation of data architecture, one that treats compliance as a continuous process rather than a one-off requirement.
By combining logical data management with open, composable systems, enterprises in the Asia-Pacific can balance innovation and control. They can harness AI responsibly, adapt to regulatory changes with confidence, and maintain measurable ROI even as technologies evolve.
The transformation underway across the region signals a broader shift in how businesses perceive data strategy. Logical data management is emerging as the backbone of modern AI readiness—a framework that unites flexibility, trust, and measurable value. In an era where agility defines competitiveness, it is becoming the cornerstone of enterprise resilience and innovation in Asia-Pacific’s digital economy.
