David Irecki, Chief Technology Officer for Asia Pacific and Japan at Boomi, discusses why many artificial intelligence (AI) initiatives stall, where the technology can realistically cut logistics costs, and what will separate AI-ready supply chains from those left behind in Southeast Asia.
Why is AI adoption in supply chains stalling beyond automation, particularly in the Philippines?
Irecki: Interest in artificial intelligence (AI) is clearly there among Philippine companies. The challenge is structural rather than technological.
Many organizations still operate with fragmented or outdated systems that are only partially digitized. The problem isn’t weak AI models—it’s the weak integration foundations underneath them.
Research from the Philippine Institute for Development Studies shows that only about 15 percent of firms in the Philippines currently use AI. Most of these are large companies in urban areas, particularly in the information and communications technology (ICT) and business process outsourcing sectors.
That leaves a large share of businesses still dependent on manual workflows.
In many supply chains, what is called a “system” is often just a combination of spreadsheets, emails and paper-based processes. That makes real-time visibility difficult. Without that visibility, AI has very little data to work with.
Data availability is another issue. The National Economic and Development Authority has noted that much of the underlying data in supply chains remains undigitized or scattered across different platforms and organizations. In many cases, the data is incomplete or outdated, particularly outside major cities.
Ultimately, AI is only as effective as the data behind it. If the data isn’t clean, connected and reliable, AI cannot move beyond simple automation into more advanced use cases such as predictive insights or autonomous decision-making.
With logistics costs estimated at about 27 percent of sales, where can AI realistically reduce costs—and where can’t it?
Irecki: AI can reduce costs significantly, but only where some level of digital maturity already exists.
In warehouses, for example, AI can help companies optimize space usage and improve demand forecasting. By analyzing historical demand patterns, organizations can avoid overstocking or running out of products—both of which create unnecessary costs.
Predictive maintenance is another valuable application. Instead of waiting for equipment to fail, AI systems can identify potential problems early. That helps reduce downtime and keep operations running efficiently.
AI can also improve operational efficiency across the supply chain—from product design and production to logistics. For instance, it can analyze operational data to identify inefficiencies and recommend better workflows.
In logistics, AI can adjust delivery routes in real time. If there is congestion at a port or delays along a transport route, shipments can be rerouted quickly. That helps reduce fuel costs and maintain delivery schedules.
These types of improvements align with the goals of the Philippines’ National AI Strategy Roadmap 2.0, which seeks to build more efficient and responsive supply chains supported by stronger governance.
However, AI is not a cure-all. It cannot build infrastructure, reduce port congestion or repair roads. Those are structural issues that require sustained investment and policy action.
AI delivers the greatest value when the basics are already in place: digitized processes, clean data and connected systems.
What are the biggest governance risks as companies connect AI across suppliers and partners?
Irecki: Once AI begins connecting multiple suppliers and partners, governance becomes significantly more complex.
This is especially true in Southeast Asia, where supply chains often span multiple countries with varying levels of digital maturity and different regulatory frameworks.
Data frequently moves across jurisdictions, each with its own cybersecurity standards, data protection rules and enforcement environments. As a result, governance cannot be designed at the level of a single company—it must consider the broader ecosystem.
One of the most immediate risks is poor data quality. If incomplete or poorly managed data feeds into AI systems, the outputs become unreliable. That can lead to poor business decisions and, in some cases, regulatory or reputational risks.
Data ownership is another challenge. When multiple partners share the same data, responsibility is not always clearly defined. Without clear accountability, problems can escalate quickly.
Security and privacy risks also increase as more systems and organizations become connected. Each connection creates another potential entry point for cyber threats, meaning a single weak link can affect the entire network.
Trust is equally important. Organizations need to understand how AI systems arrive at their decisions, particularly in regulated industries. If decisions cannot be explained, they become difficult to defend.
The solution is to build governance from the start—establishing clear data management practices, defined ownership structures and strong privacy controls before scaling AI across the ecosystem.
Agentic AI is gaining attention. What early use cases are delivering measurable results?
Irecki: Agentic AI is beginning to show value in environments where systems are connected and data is well managed.
One area where it is already proving useful is risk and compliance management. Instead of simply flagging an issue, AI agents can investigate it further—reviewing supplier histories, analyzing transaction patterns, checking contractual obligations and assessing regulatory exposure in a single workflow.
Another promising use case is disruption management. Supply chains are constantly shifting as inventories change and deliveries face unexpected delays. When disruptions occur, AI agents can quickly analyze real-time data, identify alternative suppliers and initiate corrective actions without requiring lengthy manual coordination.
This is where agentic AI differs from traditional analytics. It does not just generate insights—it can act on them, helping organizations make faster decisions and respond more effectively to disruptions.
By 2026, what will separate AI-ready supply chains from those left behind in Southeast Asia?
Irecki: The biggest differentiator will be how well companies connect and manage their data.
AI-ready organizations will have systems that communicate seamlessly with one another. Data will flow across departments—from finance and procurement to logistics—and extend to external partners. Everyone will operate from a shared, real-time view of operations.
Data quality will also be a priority. Clean and consistent data enables AI systems to produce accurate, reliable insights rather than conflicting results.
Equally important is trust. Leading companies will build transparency and accountability into their AI strategies so that decisions remain explainable rather than opaque.
Modern integration platform-as-a-service solutions such as those from Boomi help enable this environment by connecting applications, data, APIs and partner ecosystems into a unified platform. This creates a well-governed foundation for organizations to scale AI across their operations.
When that foundation is in place, AI moves beyond isolated experiments and begins to drive measurable results—from optimizing inventory and reducing operational costs to improving collaboration across supply chain partners.
Companies that achieve this level of integration will be able to scale AI across their entire supply chain. Those that do not risk remaining trapped in data silos, with fragmented systems and limited impact—and the gap between the two will only widen over time.
