ACROSS many Philippine enterprises, critical business information still sits across disconnected systems, spreadsheets, emails, and legacy platforms. While companies have invested in digital tools over the years, these systems often operate in silos, creating a significant barrier for artificial intelligence (AI) systems that rely on integrated, high-quality data. The issue is not unique to the Philippines, but it is particularly relevant in markets where many organizations are still in the middle of their digital transformation journey.
In conversation with The Manila Times (TMT), David Irecki, chief technology officer (CTO) for Asia Pacific/Japan (APJ) at Boomi, touched on why data activation is critical for enterprise AI, how integration complexity slows down AI adoption at scale, and what practical steps enterprises can take today.
THE MANILA TIMES (TMT): The first assumption in this discussion is that AI readiness is basically a data challenge. Why is data activation critical to any enterprise, especially in the Philippines?
DAVID IRECKI (Irecki): Recently, we had a change in positioning around Boomi as the data activation company. Most organizations already have a lot of data, but it's fragmented across systems. It's hard to access, and it's difficult to use in real time — which is important for AI.
Data activation is about solving that. It's all about making data discoverable and usable, not just for reporting, but for AI and real business decisions. The challenge is fragmentation. Data is spread across systems, and it is often incomplete and inconsistent.
I think that's a great use case showing that it's a data problem, not an AI problem. But the problem is that AI moves at machine speed, and only activated data can keep it secure, accurate, and useful — and without that data activation layer, AI cannot scale.
TMT: We’ve read somewhere that AI has been trained on public data, and when it comes to activating AI with data in a private environment, is that also part of the challenge?
Irecki: That falls more into the conversation around sovereignty. I think it’s almost the reverse for many businesses — they don’t want their data, which provides context to their AI models, to be out there publicly in the market.
That context is what makes their models and agents unique and enables them to solve business problems and deliver the outcomes they need within their organizations. For many of those customers, securing that data is key.
It may start with a manually curated dataset. That data can be secured through an integration layer, a traditional API layer, or newer technologies like MCP, to ensure that those AI agents and models only access the data they are permitted to when operating within the business.
TMT: As companies activate more data in AI, how do data management controls — like governance, visibility, and security — become critical in the integration process, especially with recent findings that insider threats are growing alongside external threats?
Irecki: As enterprises activate more data across the organization, governance becomes essential. In the Philippines, the National Privacy Commission’s AI advisory already requires businesses to have data governance and accountability in place.
With AI agents, the stakes are even higher because they don’t just read data — they act on it in real time. For a business, that means you need to focus on three things: visibility into how data is being used, control over what actions are allowed (both for the agent and data access), and full auditability of every decision.
If you’re giving these agents autonomy without that level of orchestration, it creates significant risk. But giving agents autonomy with strong governance is actually a powerful advantage. That’s where the data activation layer comes in — it ensures that data is trusted, governed, and used responsibly.
TMT: In that context, how do AI-driven threats from the outside, for instance, affect the integrity of data pipelines?
Irecki: AI introduces a new attack surface because those agents interact with multiple systems. We’re already seeing use cases like data poisoning and over-permissioned access becoming more serious risks.
If an agent is working with compromised or low-quality data, it will make incorrect decisions — and do so at scale. Security and resilience across data pipelines — through controlled access, continuous monitoring, and trust — have to be foundational from the very beginning, not an afterthought.
AI is already embedded in the business, so if you don’t build security in early, you risk scaling those vulnerabilities instead of solving them.
TMT: In that push for AI-driven data activation, it appears that the focus is more on low- to mid-level supervisors of data and data management. In that context, how do government and private sector industry leaders need to contribute to raising awareness about data activation?
Irecki: It’s a leadership issue as much as it is a technology issue. Leaders within an enterprise need to prioritize integration and data governance as core parts of their AI strategy.
In the Philippines, we’re definitely seeing strong government support, with the Department of Science and Technology investing P2.6 billion in AI projects, and the National AI Strategy continuing to evolve alongside that.
But enterprise leaders need to match what the government is driving by connecting their systems, standardizing their data, and embedding governance early in those processes. Integration and data governance are no longer just technical challenges — they’re strategic capabilities.
TMT: How does the Boomi platform contribute to data integration in the country?
Irecki: Boomi is the data activation company, and it helps solve fragmentation across systems, improve the quality of that data, and enable real-time use for AI agents.
That shift is critical — businesses move from using data only for reporting to using it for actionable insights driven by AI, which is increasingly important because access to trusted data directly impacts business outcomes.
