AI Is Moving from Interface to Infrastructure
- Sertis
- Mar 23
- 10 min read
AI has become a regular topic in leadership meetings, strategy discussions, and digital transformation plans. But for many organizations today, the real question is no longer whether they should use AI.
The more important question is this: How can AI create real business results?
That was the focus of a recent exchange during the Sasin Turbo Program office visiting, where participants explored a bigger view of AI’s role in business. The session was led by Aubin Samacoits, Chief AI Officer at Sertis, who shared a perspective that went beyond headlines and hype. Instead of looking at AI as just another fast-moving technology trend, the discussion centered on something far more important for business leaders: how AI can help organizations make better decisions, work more effectively across systems, and turn strategy into action.
That shift matters, because the AI conversation is changing quickly. For a while, many people saw AI mainly as a tool for writing, summarizing, or answering questions. Useful, yes, but still limited. Today, that view is evolving. Businesses are starting to look at AI differently, not just as an assistant, but as a real operational force that can help teams analyze information, connect workflows, speed up execution, and improve decision-making.
This is why AI is no longer just about adopting a new tool. It is about redesigning how work gets done. And that is exactly why ideas like Agentic AI and data-driven transformation are becoming more meaningful for organizations that are not interested in technology for its own sake, but in outcomes that can actually be measured.
The Question Is No Longer “Do We Have AI Yet?”
Over the past year, many organizations have already started using AI in one form or another. Some teams use it to summarize information. Some use it to speed up writing. Others use it to search internal knowledge or make everyday tasks move faster.
But once AI becomes something that almost every organization can access, the real question changes.
What separates companies that are simply experimenting with AI from those that are already getting business value from it is not whether they have the tool. It is whether AI has been placed in the right part of the business and whether it is actually helping important work move forward.
Because in business, the speed of getting an answer is not the same as the speed of making a decision. And it is definitely not the same as the speed of execution.
That is why the AI conversation is moving beyond innovation for innovation’s sake. More and more, it is becoming a conversation about competitive advantage.
AI Is Growing Fast. Business Value Is Not.
The market is sending a clear signal: AI is no longer a passing trend.
Today, 78% of organizations globally are already using AI. Investment in generative AI has reached US$33.9 billion, and Southeast Asia’s digital economy is projected to reach US$300 billion in GMV by 2025. At the same time, only 39% of organizations say AI is making a clear impact on EBIT at the enterprise level.
That contrast says a lot.
It shows that access to AI does not automatically translate into business value. Many companies already have tools. Many already have pilots. Many already have use cases. But not all have reached the point where AI is meaningfully reducing cost, improving decisions, or changing how work operates across the business.
That is where the real gap is.
In the early stage, AI often creates value through modest productivity gains. It helps people work faster, write faster, search faster. But once organizations move toward AI that understands goals, works across multiple steps, and supports execution beyond a single prompt, the opportunity becomes much bigger. At that point, the value is no longer just about individual productivity. It starts affecting the throughput of the wider system and, in some cases, opens the door to 40%+ operational cost reduction.
The Real Shift: AI Must Help Work Move Forward
For many businesses, AI has so far been something reactive. A user asks, the system answers. It may generate text, summarize a report, write code, or return an image. But the broader direction of AI in 2025 and 2026 is clearly moving beyond that model.
Businesses are starting to look for AI that can understand an objective, think through the next steps, use tools, and help work continue toward a result.
That is where Agentic AI becomes important.
The value of Agentic AI is not that it sounds more advanced. The value is that it moves AI from being a system that responds well to a system that can help work get done. It is the difference between an assistant that gives a smart answer and a system that helps teams actually progress.
Another important shift is this: competitive advantage is no longer only about the model itself. It is increasingly about workflow design, tool orchestration, system connectivity, and how well data moves across the organization.
In other words, businesses do not just need AI that sounds smarter. They need operations that work smarter.
AI Can Only Go as Far as Your Data Allows
No matter how advanced AI becomes, it will struggle to create real business impact if organizational data is still fragmented, stuck in silos, or spread across disconnected systems.
This is why data-driven transformation is no longer a back-end issue. It is now one of the foundations of long-term business advantage.
Organizations that treat data as a long-term asset, and manage it as reliable, reusable infrastructure, are usually able to scale AI faster than organizations that prepare data one project at a time. The difference is not just technical maturity. It is a strategic discipline.
This also reflects a broader truth about successful AI transformation. The organizations that move further tend to start with a real business problem, give as much importance to people as to technology, treat data as infrastructure, work across teams early, and build governance from the beginning.
That is also why many companies are now shifting more of their AI budgets away from licenses alone and toward data organization and system integration. Because for AI to create real value, it cannot only be generally intelligent. It has to understand the business it is working inside.
Southeast Asia Is Not Watching from the Sidelines
Southeast Asia is no longer at the stage of simply observing AI from a distance. The region is already in an acceleration phase.
Many organizations are moving from piloting use cases to scaling them. That matters because the leadership question in this region is changing too. It is no longer “Should we start using AI?” It is “How do we make AI work across the organization in a way that is practical, scalable, and reliable?”
As the region’s digital economy continues to grow, AI decisions are becoming more than innovation signals. They are becoming part of business competitiveness.
Why Sovereign AI Matters More Here
Another major shift is the rise of Sovereign AI: AI development that better reflects local infrastructure, in-country processing, data governance requirements, and language realities.
This matters especially in Southeast Asia, where language, behavior, and regulatory contexts differ significantly across markets.
For businesses operating in this region, AI cannot be treated as one-size-fits-all. Local language support, local context, and stronger control over data are becoming increasingly important. That is why investment in localized infrastructure, sovereign LLMs, and systems that support languages like Thai and Bahasa is becoming more visible.
This is not only about national strategy. It is also about business usability. AI becomes more valuable when it understands the environment it operates in.
The Better Question Is Not “How Can We Build a Better Chatbot?”
From a business point of view, the better question may be: how can we remove unnecessary steps from work?
Organizations that use AI only as a better question-and-answer interface often see gradual gains. But organizations that use AI to take over parts of a workflow, coordinate across systems, and help teams move faster across multiple steps tend to see a very different level of value.
This is the logic behind an agent-ready organization. It is not just about having the technology. It is about having cleaner data, connected systems, updated measurement models, and teams that are ready to shift from operator roles into orchestrator roles.
The goal is not simply to build a better chatbot. The goal is to build a digital worker that can reduce manual effort and improve execution.
Four Industries Where the Opportunity Is Becoming Clearer
When you look across industries, the pattern is strikingly similar. There is more data, more complexity, and more pressure to act faster.
That is why the opportunity for AI is no longer just about helping businesses see information more clearly. It is about helping teams interpret faster, coordinate better, and act sooner.
Retail: Moving beyond reporting and into areas like assortment decisions, demand sensing, promotion optimization, and faster commercial action.
Finance: Becoming more relevant in fraud monitoring, risk workflows, compliance support, and knowledge-intensive decision-making.
Manufacturing: Helping improve predictive maintenance, quality intelligence, process visibility, and operational response.
Energy: The focus is increasingly on forecasting, asset intelligence, operational prioritization, and resilience planning.
Across all of these sectors, the new value of AI is not just in analysis. It is in helping businesses move from information to action.
Real Use Cases Are Making the Business Value Easier to See
The strongest proof points for AI are no longer theoretical.
In manufacturing, the case of Midea Thailand shows how 5G and AI inspection can reduce rework by 75%, improve production efficiency, and contribute to recognition as Thailand’s first Global Lighthouse Factory by the World Economic Forum in 2025.
In logistics, the case of WHA Group Thailand shows how real-time fleet visibility, data-driven route planning, predictive battery monitoring, and automated ESG reporting can improve operational visibility, reduce cost, and support the scale-up of EV logistics.
In property technology, the Nestopa AI Agent demonstrates how AI can transform property search from a rigid filter-based experience across more than 130,000 listings into a more intelligent, multilingual search experience with context memory. The result is stronger engagement, a 70% increase in agent productivity, and better access to international markets.
In healthcare, AI-assisted chest X-ray screening developed by Siriraj Hospital and its partners shows how hospitals with heavy workloads and limited radiology capacity can screen more consistently, identify high-risk cases faster, and scale support across healthcare networks.
In banking, a Knowledge Management AI system supporting 1,500 users across multiple teams has helped improve productivity, enhance branch sales processes, increase conversion rates by 12%, and reduce both mis-selling risk and internal knowledge gaps.
Taken together, these examples make one thing clear: the real business value of AI does not come from AI being impressive. It comes from AI being useful inside real operations.
From Vision to Execution: Real AI Needs Structure and Clear KPIs
One reason many organizations still struggle to create measurable value from AI is not that the technology is missing. It is that the business design is still unclear.
If it is not clear where AI should support the business, how it should work with existing systems and data, and how success should be measured, AI often remains stuck in the category of “interesting” instead of becoming a real business capability.
In practice, enterprise AI needs to begin with a clear goal. What should AI help improve? Which decisions should it support? Which part of the workflow should it accelerate? Where should it reduce burden?
From there, AI needs access to reliable knowledge and trusted data. Without that layer, its output may sound good, but it will not be grounded in the real context of the organization.
Once the goal and knowledge layer are in place, AI can move into reasoning, sequencing, and execution. That is when it starts helping work move forward inside actual business systems rather than stopping at the chat interface.
A practical AI agent structure usually starts with Goal and Context, followed by Knowledge and Memory, then Reasoning and Planning, and finally Action and Integration. Around all of it, there must be Governance and Trust through guardrails, human oversight, auditability, and control.
A Simple Way to Think About the Architecture
At a technology level, this kind of AI often works through three main layers:
Perception Layer
Where the system receives and gathers input from multiple sources
Cognition Layer
Where it interprets context, reasons, and decides what to do
Action Layer
Where it connects to execution through service integration, monitoring, and feedback loops
Three KPIs That Matter Most
There are many ways to evaluate AI, but three metrics remain especially important:
Cost Reduction
Is AI helping lower costs in a meaningful way?
Time to Completion
Is work getting done faster?
Error Rate
Is AI reducing mistakes and inconsistency?
The More Ready Organizations Are Not Just More Advanced. They Are More Disciplined
Many organizations want to go further with AI. But readiness is not only about how much technology they have bought.
It is also about whether their data is organized, whether their systems connect efficiently, whether their metrics support long-term business value, and whether their teams are ready to evolve from doing work alone to working effectively with AI.
Organizations that succeed in AI transformation usually follow the same core principles. They start with real business problems. They treat people and technology with equal importance. They view data as a long-term asset. They work across teams early. And they build governance and security from the beginning.
That is the real difference between doing an AI project and building a new business capability.
What Businesses Should Take Away
From here on, AI will not be judged only by how well it answers questions. It will be judged by how well it fits into real work.
The organizations that gain an advantage will not always be the ones with the most advanced technology. Often, they will be the ones that understand their own data best, design their workflows more clearly, and know exactly where AI should be placed to create the clearest business outcome.
When AI is applied in the right place, it stops being just a new tool. It becomes a new way to move the business forward.
FAQ
What is Agentic AI?
Agentic AI refers to AI that goes beyond answering questions or generating content. It can take a goal, think through multiple steps, use tools, and help work move forward more independently.
Why is data-driven transformation important?
Because AI only creates real business outcomes when it can work with reliable, connected, and usable data. If data is fragmented, AI may be useful on screen but limited in real operations.
How should organizations measure AI success?
A practical framework starts with three core metrics: cost reduction, time to completion, and error rate.
Which industries are seeing strong AI opportunities?
Retail, finance, manufacturing, energy, logistics, real estate, healthcare, and banking are all seeing growing AI potential. What they share is the need to manage more data, more complexity, and faster decision-making.
Why is Southeast Asia an important AI region to watch?
Because the region is no longer experimenting at the edges. It is moving into real enterprise execution, supported by digital growth, infrastructure investment, and stronger interest in localized and sovereign AI models.
As AI moves beyond answering questions and becomes part of how businesses decide and act, the challenge is no longer just where to start. It is how to design AI in a way that connects with real data, real systems, and real business goals. For organizations looking to turn AI from a promising idea into measurable business impact, building the right foundation now matters more than ever.
