Southeast Asia is one of the fastest-growing digital economies in the world. According to Kearney, AI is expected to contribute nearly US$1 trillion to the region’s GDP by 2030, reflecting a growing appetite among governments and enterprises to adopt intelligent technologies. From automated customer service to smart logistics, the potential use cases for AI are rapidly expanding across the region.
Yet, despite this optimism, enterprise AI readiness in the region remains limited. Cisco’s 2024 AI Readiness Index shows that only 15% of businesses in Asia Pacific, Japan, and China (APJC) are prepared to scale AI initiatives. That’s a sobering figure for a region with such ambitious digital transformation goals.
Infrastructure is one of the key obstacles. AI requires much more than algorithms and machine learning models; it depends heavily on data infrastructure. Most organisations still operate within siloed systems, where different departments, tools, or business units cannot communicate efficiently. These fragmented environments are poor foundations for real-time data processing AI needs.
Based on same Cisco study, 82% [of enterprises in the region] are still grappling with fragmented data systems that limit access to the real-time insights AI needs to deliver value. The situation calls for a hard look at the architectural backbone that supports enterprise AI, not just the AI layer itself.
AI success in Southeast Asia hinges on ambition and a clear understanding of what’s under the hood. To move beyond pilot programmes and proof-of-concepts, enterprises need to build infrastructure that is agile, responsive, and resilient by design.
The integration gap is AI’s weakest link
AI has evolved beyond static models or standalone functions. Today’s agentic AI systems, those capable of reasoning, acting autonomously, and learning in context, require constant streams of high-quality data. They cannot afford delays, gaps, or incomplete context. However, many organisations in Southeast Asia aren’t ready for this level of integration.
“Today’s AI is no longer static or standalone. Agentic AI systems, unlike earlier generative tools, are dynamic, distributed, and inherently intertwined with real-time data to deliver meaningful outcomes,” says Floyd Davis, Vice President of Solution Engineering for APJ & ME at Solace. Real-time data availability isn’t optional for these systems to function—it’s fundamental.
The reality, however, is sobering. 81% of IT leaders in APAC cite data integration as one of the biggest challenges to deploying AI effectively. This isn’t just a technical issue; it’s an architectural one. In a typical enterprise, data is scattered across SaaS apps, internal databases, legacy platforms, and third-party systems, none of which were built for seamless interoperability.
The result is a breakdown in context. Imagine a fraud detection system that doesn’t receive data about a recent transaction in time, or a customer service bot that isn’t aware of an ongoing issue from another channel. AI delivers subpar results without complete and connected data, creating user frustration, regulatory risks, or missed opportunities.
Davis argues that the AI conversation needs to shift toward digital resilience: “It requires an underlying, AI-ready architecture platform capable of real-time data integration and distribution. Without this foundation, businesses risk fragmentation, latency, and limited adaptability to shifting market dynamics.”
Why event-driven architecture is foundational
This is where event-driven architecture (EDA) becomes a game-changer. Traditional request-response models treat data like a transaction, waiting for a system to ask for something before responding. EDA flips that model. It enables systems to listen for changes (events) and react instantly across distributed environments.
In Davis’s words, “An EDA platform enables systems to detect, process, and respond to real-time events the moment they occur—whether it’s a point-of-sale transaction, a shipment delay, or a customer inquiry.” This model powers the kind of responsiveness and agility that agentic AI systems require.
At the heart of EDA is the event mesh, a layer that connects different systems and data sources, allowing them to share events as they happen. This mesh functions like a digital nervous system, routing real-time information to wherever it’s needed, whether in the cloud, on-premise, or across multiple geographies.
In practical terms, this could mean a delivery platform rerouting drivers in response to real-time traffic, or a bank flagging suspicious activity within milliseconds. These aren’t futuristic use cases as they’re happening today in enterprises investing in real-time architectures.
Solace, a leader in this space, helps enterprises deploy event mesh architectures that deliver “the timely, contextual information [AI] needs to make accurate decisions at speed and scale.” Their infrastructure underpins critical finance, logistics, and telecommunications systems across APAC.
Agent mesh gives AI the intelligence to act
If the event mesh is the nervous system, then the agent mesh is the brain. This second layer—also championed by Solace—adds autonomous, intelligent agents into the system. These agents aren’t just passive listeners; they can reason, learn, and act based on real-time data.
As Davis explains, “True resilience requires more than speed; it demands systems that can think, adapt, and act independently. This is where the agent mesh comes in.” Agent mesh introduces a layer of intelligence that works in harmony with the data flow enabled by the event mesh.
These agents can operate autonomously or with human oversight. For example, one agent may monitor user interactions for signs of friction, while another may adjust system configurations in response to observed performance trends. Together, they orchestrate a fluid, adaptive enterprise that responds to predictable patterns and unexpected disruptions.
“Think of it as adding a layer of distributed intelligence to your enterprise nervous system,” says Davis. This analogy captures the value of an integrated architecture that doesn’t just transmit data but also understands and acts on it.
Solace’s approach with agent mesh is not a theoretical model. Enterprises using their platform already deploy AI-powered agents to automate decisions, detect anomalies, and streamline operations. The dual-layered architecture, event mesh plus agent mesh, is what Davis calls the “dual engine of resilience.”
What Southeast Asian enterprises must do next?
The AI revolution is accelerating. But the winners won’t just be those who build the smartest models. They’ll have the most resilient foundations that can scale AI across business functions, geographies, and use cases without crumbling under complexity.
That means rethinking architecture. Enterprises should invest in real-time, event-driven systems instead of relying on static integrations and brittle APIs. They should introduce distributed agents that can act autonomously and intelligently. And they should modernise their infrastructure with future-proof platforms like those offered by Solace.
“In an AI-first world, digital resilience is no longer a nice-to-have; it is a strategic imperative,” says Davis. “Organisations must build systems that are not only intelligent but also agile, adaptive, and resilient by design.”
The formula is clear. AI maturity isn’t just about the model; it’s 20% algorithms and 80% data and integration. Enterprises that get their foundations right will be better positioned to capture the full promise of AI, whether it’s serving customers, defending against threats, or driving innovation.
To thrive in the next wave of digital transformation, Southeast Asia’s enterprises must not only do AI but also become AI-ready at their core.
