For ASEAN CIOs in 2026, the AI paradox is clear: budgets are rising, yet financial returns are not. With almost half the region scaling AI beyond pilots, we have hit a “data wall“. The shift toward agentic AI has only intensified the pressure, transforming integration from a back-office concern into the new control plane for security, privacy, and the execution of sovereign AI.
To move from ambition to monetisation, leaders must stop managing data in silos and start activating it in real-time across finance, supply chains, and customer systems.
Without a unified data fabric that ensures governance and trust, autonomous agents cannot operate safely at scale. Data activation is no longer an IT project; it is the core strategy for turning AI spend into EBIT growth.
As David Irecki, chief technology officer for Asia Pacific and Japan at Boomi, observes: “AI has outpaced enterprise data readiness.” The technology is ready; the question is whether enterprise foundations are.
The data wall: where ASEAN CIOs are getting stuck
Despite strong adoption momentum, the financial impact of AI is not following. McKinsey data cited in a Singapore Economic Development Board report reveal that 60% of organisations see an impact of less than 5% on EBIT from AI initiatives, with nearly one in five reporting no measurable return.
According to Irecki, CIOs are getting stuck in three critical places.
“First is around data quality, because if you have multiple systems that don’t agree on what the customer or the order is, that’s going to cause your AI to hallucinate. Second, it’s around integration debt. All those years of point-to-point connections can’t scale in this new world of AI. And third, governance. Organisations don’t just fully trust the data behind those AI decisions.” David Irecki
This governance gap is acute. A Boomi and FT Longitude report found that just 2% of organisations have fully accountable AI agents, while nearly 80% lack visibility or control over agent behaviour. Meanwhile, MIT’s “State of AI in Business 2025” notes that 90% of employees use AI tools informally, yet only 40% of organisations officially support them—creating unmanaged operational risk.
From lakes to live streams: what data activation really means
Beyond data lakes and static dashboards, “data activation” represents a fundamental shift. “For us, data activation is all about a shift from storing data to using it in real time,” explains Irecki. “Today’s AI agents need live trusted data that they can act upon in real-time inside workflows running within your business.”
In practice, this means connecting ERP to CRM to supply chain, finance, and support systems; exposing that data through governed APIs; and ensuring it is contextual, consistent, and policy-compliant. The payoff is tangible: instead of static reports, enterprises gain real-time, actionable decisions.
Analyst projections underscore the urgency. By 2027, an estimated 80% of agentic AI use cases will depend on real-time access to contextual data to deliver ROI—a threshold many ASEAN enterprises have yet to meet.
APIs as the new control plane for sovereign AI
As agentic AI explodes in 2026, APIs and integration are emerging as the essential control plane for security, privacy, and sovereign AI. “APIs and integration become that control plane,” says Irecki. “By routing all of your AI activity through governed APIs, you’re able to enforce data access policies, ensure masking or compliance rules are applied, and audit every decision.”
This architecture is particularly critical in a region where digital sovereignty is reshaping infrastructure choices. Forrester predicts that sovereignty will shape AI infrastructure decisions for half of APAC firms in 2026, as nations assert digital autonomy through data residency requirements and regional provider preferences.
For ASEAN CIOs, this means building integration layers that can enforce sovereign boundaries while enabling cross-border innovation—a delicate balance that demands both technical rigour and policy fluency.
Governing the edge: balancing speed with compliance
With Boomi reporting that 90% of staff use AI informally, the challenge is not whether to enable experimentation, but how to channel it responsibly. “You can’t stop it. You must be able to channel it in some way,“ Irecki advises.
“Instead of embedding governance in every application, you can centralise it. You can mask sensitive data before it reaches AI, route different types of data to different models, and enforce access rules automatically.” David Irecki
This approach allows organisations to maintain speed and adoption while preserving control—a critical capability as regulatory frameworks across ASEAN continue to evolve.
Singapore’s Monetary Authority of Singapore, for instance, has issued guidelines emphasising fairness, explainability, and auditable decision-making for AI in financial services. Similar momentum is building across the region, making embedded governance a strategic imperative rather than an afterthought.
Killing data debt: the business case for integration-centric architecture
What exactly is “data debt“? For Irecki, it is the accumulation of scripts, manual processes, and disconnected systems that IT teams inherit while under pressure to deliver at speed, not scale. “Over time, that becomes a hidden tax on the organisation. It shows up as broken integrations, conflicting reports, slow project delivery, and AI initiatives that stall,” he elaborates matter-of-factly.
The business impact is threefold: cost (time spent fixing issues), risk (bad data driving decisions), and opportunity (delayed AI value realisation). PwC research in Singapore highlights the scale of the challenge: only 37% of organisations report having a single trusted data record, compared with 59% among recognised AI leaders. Those with stronger data foundations report seven times the revenue and efficiency gains.
Building a business case to eliminate data debt requires shifting the conversation from IT maintenance to strategic enablement. Reusable, governed integrations are not a cost centre; they are the foundation for scalable AI value.
Fastest payback: Where data activation delivers ROI first
Where should ASEAN CIOs focus for the quickest returns? According to Irecki, the pattern is consistent: “The fastest return is in those high-volume, repeatable processes.” Customer service often delivers early wins by improving call routing and reducing resolution times.
Sales teams gain from deeper customer insights and improved cross-sell. Supply chain benefits from better forecasting and faster exception handling. Finance sees value in improved credit decisions and accelerated collections.
“The pattern’s consistent that data activation delivers value fastest where those decisions are frequent and they’re directly tied to revenue or an improvement in customer experience,” Irecki notes. This pragmatic, use-case-led approach helps CIOs demonstrate tangible ROI while building the architectural foundation for broader AI scale.
Three non-negotiables for turning AI spend into EBIT by 2026
If Asia has hit a data wall, what are the non-negotiables for breaking through? Irecki offers three clear priorities:
- A reliable data foundation: consistent definitions, strong data quality, and robust governance.
- An integration-centric architecture: APIs as the primary interface for AI access to back-end systems.
- Embedded governance: security, privacy, and compliance built into workflows, not added later.
“Having those three foundations within their business, they’re seeing seven times the revenue and efficiency gains,” Irecki observes. For ASEAN CIOs, this is not merely a technical checklist; it is a strategic roadmap for converting AI ambition into measurable business outcomes.
As the region moves deeper into the age of agentic AI, the organisations that lead will be those that treat data activation not as an IT project, but as the core strategy for turning AI spend into EBIT. The technology is ready. The question now is whether enterprise foundations are.
Click the PodChats player for in-depth discourse with Irecki.
- Boomi cites an interesting statistic: 95% of organisations struggle with AI ROI, yet you argue the models are ready. Where are ASEAN CIOs getting stuck—and why are data quality and integration now the real villains, not the AI? (Singapore report)
- Beyond lakes and dashboards, what does “data activation” mean in practice? How do we put live, trusted data into the flows AI agents rely on?
- As Agentic AI explodes in 2026, why should CIOs see APIs and integration as the new “control plane” for security, privacy, and sovereign AI?
- According to Boomi, with 90% of staff using AI informally, how do we enforce data governance without killing speed? Can the integration layer automate policy enforcement?
- Define data debt. How do we build a business case for killing “data debt” and moving from brittle scripts to reusable, AI-ready integrations?
- According to Salesforce, 96% of CIOs say success hinges on integrating AI into the day-to-day workflow. Where do ASEAN CIOs see the fastest payback for data activation—finance, supply chain, or customer service?
- How does activating data for a swarm of autonomous agents differ from powering a simple chatbot? How do we make agents from different vendors talk to each other?
- If Asia has hit a data wall, what are the three non-negotiables for CIOs to turn AI spend into EBIT (Earnings Before Interest and Taxes) by 2026?
- Any takeaways for CIOs on how to shift their CIO initiatives from delivering disappointing results to powering business innovation?
