The shift toward contextual AI and decision intelligence in Southeast Asia

A decade ago, the challenge for businesses was gathering data. Today, the issue is no longer scarcity, but relevance. Across the region, companies are dealing with massive volumes of information yet often struggle to convert that data into meaningful, actionable insights. This disconnect between collection and application underscores a deeper tension: many organizations are still navigating the complex journey from digital transformation to data maturity.

This evolution has given rise to a new imperative—building analytical cultures rather than just data teams. The growth of data-intensive businesses, from fintech to logistics and urban infrastructure, has triggered demand for more than just dashboards or predictive models. The focus is increasingly on embedding data into the very fabric of decision-making processes, where traditional analytics and emerging tools like generative AI work in tandem.

---

---

In markets like Singapore, where regulatory frameworks such as the AI Verify Foundation and the Data Protection Trustmark are in place, companies are experimenting with AI not as a replacement but as an augmentation layer that enhances both customer experience and operational precision. In contrast, in emerging economies within the region, the emphasis often lies in leapfrogging legacy systems—bypassing intermediate tech stacks and moving straight to hybrid, scalable solutions.

To better understand how things are evolving, we speak to Joanna Teo, Managing Director of Attribute Data. Recognised as an honoree in IMDA’s 2023 SG Top 100 Women in Tech, Joanna has been at the forefront of the data industry in the region.

What makes Southeast Asia particularly compelling isn’t just its diversity of digital readiness, but the regional nuances that shape how data is used. In Thailand, business hierarchy influences data adoption patterns. In Indonesia, hyper-localisation drives platform development, especially in sectors like ride-hailing and e-commerce. In China, pragmatism drives speed and scalability—an insight now informing how startups and investors approach regional expansion strategies.

This has important implications for both startups and investors. It’s no longer enough to build AI-powered platforms; the real differentiator lies in how well these technologies are grounded in domain expertise, measurement discipline, and regulatory foresight. The most forward-thinking players are not trying to sidestep regulations—they’re integrating compliance into product design. This mindset—where trust and functionality co-exist—defines the new generation of data-first companies emerging from Southeast Asia.

As regional economies adapt to new models of decision automation and real-time analytics, the winners will be those who understand that the real innovation lies not in the novelty of the technology but in its contextual application. Southeast Asia is not simply catching up to global trends—it’s rewriting how data, AI, and strategy coalesce in one of the most dynamic regions in the world.

You wear multiple hats at Attribute Data, from founder to head of data intelligence. What problem were you originally trying to solve when you started the company, and how has that evolved with today’s data needs?

When I founded Attribute Data, I was struck by a paradox that kept appearing across organizations in Southeast Asia: companies were drowning in data yet starving for insights. It reminded me of what psychologists call the “paradox of choice” – when presented with too many options, we often make worse decisions or no decision at all.

The original problem I aimed to solve wasn’t just about gathering more data – businesses had plenty of that. It was about identifying which attributes actually matter for decision-making. In Singapore and across the region, I noticed organizations collecting massive datasets without the fundamental ability to extract meaningful signals from the noise.

What’s interesting is how this challenge has evolved. Initially, we focused on helping clients build data foundations and basic analytics capabilities. Today, the conversation has shifted dramatically. Now we’re engaged in what I call “decision engineering” – helping organizations not just understand their data but deploy it as a strategic asset that drives concrete actions across sectors from government ministries to financial institutions.

The evolution mirrors Southeast Asia’s digital transformation journey. Where once we helped clients ask “what happened?”, we now help them model “what might happen if?” This shift from descriptive to predictive and prescriptive analytics reflects the region’s growing data maturity. Singapore may lead this curve, but what excites me is seeing similar transformations happening at different paces across Thailand, Vietnam, Indonesia, and China, each with their unique cultural contexts shaping how data is perceived and utilized.

Our recent expansion into China, as they reopened after COVID-19 ,has revealed a distinctly pragmatic approach to data utilization. Chinese organizations often prioritize speed and scalability in implementation, with less emphasis on theoretical perfection. In Thailand, we’ve observed how the hierarchical business culture influences data communication – insights often need to be presented through traditional authority channels to gain acceptance. Meanwhile, in Indonesia, the diverse archipelagic nature of the market has created unique data challenges where companies like Gojek and Tokopedia have pioneered hyper-localized analytics that account for dramatic regional differences in digital adoption and consumer behavior.

With the explosion of interest in generative AI, how do you see the role of traditional data analytics shifting, especially in industries like finance and logistics where precision is key?

There’s a moment I often share with skeptical executives about the relationship between generative AI and traditional analytics. I ask them to imagine a master chef in a kitchen. Generative AI is like giving that chef a revolutionary new cooking technique, but traditional analytics remains the ingredients, the recipe, and the knowledge of how flavors work together. You need both to create something extraordinary.

In precision-driven industries like finance and logistics, this relationship becomes even more crucial. Traditional analytics provides the guardrails and foundations that keep generative AI from veering into fantasy. DHL Express exemplifies this approach, recognizing AI as a strategic enabler rather than a replacement for traditional analytics. Their advanced analytics capabilities, powered by AI, optimize routes, predict shipment volumes, and proactively manage potential disruptions. This integration translates to faster transit times, reduced costs, and improved reliability for customers while maintaining the precision logistics demands. They’re also exploring AI-powered robots and drones for warehouse management and last-mile delivery—innovations that require traditional analytical foundations to function effectively in complex real-world environments.

In the financial sector, organizations like Singapore’s DBS Bank have pioneered this hybrid approach. The POSB digibank application has integrated traditional credit scoring models with conversational AI to create more personalized financial planning and wealth management recommendations in the form of content to provide customers with an added layer of insight of their own transactional history. Traditional analytics ensures regulatory compliance and risk management precision, while generative capabilities enhance customer engagement and personalization.

What’s unique to Southeast Asia is the uneven digital infrastructure across the region. In Singapore or parts of Malaysia, organizations can leverage generative AI’s full potential with robust data pipelines. Meanwhile, in emerging markets, we’re seeing hybrid approaches where generative AI compensates for data gaps while traditional analytics ensures accuracy where data is strong.

The most successful organizations don’t see this as an either/or proposition. They’re creating what I call “analytical ecosystems that augment decision making”, where traditional methodologies and generative capabilities complement each other, creating a whole greater than the sum of its parts.

Many organizations struggle to move from data collection to actual decision-making. What’s the most common gap you’ve observed between data teams and business teams—and how do you bridge it?

This reminds me of the “false friends” or false cognates phenomenon in linguistics, when two words in different languages look similar but have entirely different meanings. Data scientists say “significant” and mean “statistically valid,” while business leaders hear “important.” Everyone nods in agreement while fundamentally misunderstanding each other.

The most common gap I’ve observed across Southeast Asian organizations isn’t technical – it’s translational. Data teams speak in confidence intervals and p-values; business teams speak in revenue growth and customer satisfaction. Even when they’re discussing the same objective, they’re using fundamentally different languages to describe it.

At Attribute Data, we’ve developed what we call the “Data Value Path” – a four-step process that transforms raw data into strategic assets. This framework evolved from our direct experience with the challenges organizations face in bridging the data-to-decision gap.

The first step is “Identify What Data You Need” – auditing existing data resources, conducting requirements workshops, and establishing measurement frameworks that align with business objectives. We helped a financial services client train subject matter experts who could work as translators between technical and business teams, identifying the metrics that truly drive business outcomes rather than what’s merely available.

The second step, “Fix Your Data Problems,” addresses structural issues – consolidating silos, developing coherent data architecture, and establishing governance frameworks. The third step, “Make Sense of Your Data,” focuses on developing the intelligence needed for insight generation through appropriate pipelines and visualization.

Most critically, our fourth step, “Augment Your Decision Making,” focuses on embedding data capabilities directly into decision processes through expert recommendations, experimental design, and data enrichment that accelerates customer-centric innovation.

What’s particularly interesting in Southeast Asia is how these gaps are colored by cultural factors. In hierarchical organizational cultures common in parts of the region, junior data analysts may hesitate to challenge senior business leaders’ assumptions, even when the data contradicts conventional wisdom. Creating safe spaces for data-driven dialogue becomes as important as the technical analysis itself.

The organizations that successfully bridge this gap don’t just get better data – they fundamentally transform how decisions are made, creating cultures where evidence and experience reinforce rather than compete with each other.

You’re also active in venture capital with Protege Ventures. From your perspective, what makes a data-driven startup stand out today, beyond just having an AI label?

We tend to see objects only in their traditional roles, a phenomenon called “functional fixedness” . In venture capital, particularly around data and AI startups, there’s a similar fixation. Too many founders are hammering away at the same handful of problems with increasingly sophisticated technical hammers. Yes, we kept hearing: “I’m an AI startup”.

What truly makes a data-driven startup stand out is when it breaks this fixedness. The most compelling companies I’ve evaluated through Protege Ventures aren’t just applying sophisticated algorithms to traditional problems – they’re reimagining what problems can be solved in the first place.

Take ZOLO, the AI-powered B2B SaaS company I co-led an investment in, which streamlines order management for food suppliers. Their brilliance wasn’t in having marginally better algorithms than competitors. It was in recognizing that the food supply chain in Southeast Asia operates fundamentally differently than in Western markets, with unique informal relationships, payment preferences, and logistical challenges. They built their data model around these regional realities rather than importing Silicon Valley assumptions.

Beyond specific use cases, three characteristics consistently distinguish exceptional data-driven startups from the merely technically competent.

First, they demonstrate domain mastery – deep understanding of the industry they’re disrupting that shapes how they collect, analyze, and apply data. This domain mastery is exemplified by a Singapore-based space tech startup, Equatorial Space Systems, I recently evaluated that develops hybrid-engine rockets and space launch services. They’re giving SpaceX meaningful competition by recognizing the unmet demand for satellite launches. What impressed me wasn’t just their technical capabilities but their founder’s clear-eyed understanding of the market dynamics and regulatory landscape. He maintains a laser focus on developing lower-cost, more environmentally sustainable jet engines and fuel while keeping burn rates remarkably low.

Second, standout startups show measurement discipline – they’re rigorous about defining success metrics that connect directly to customer value, not just technical performance. Both startups’ founders’ measurement discipline is evident in the regular, focused updates they provide investors, sharing metrics that truly matter rather than vanity statistics. He understands that a deep tech startup is fundamentally in survival mode and has attracted highly competent talent aligned with that reality.

And finally, they embrace ethical foresight – anticipating how their data usage might impact various stakeholders over time, particularly in Southeast Asia’s rapidly evolving regulatory landscape.

The region presents unique opportunities for data-driven startups precisely because of its diversity. A recommendation algorithm optimized for Singapore’s highly digital population might fail completely in Indonesia’s tier-two cities. The startups that recognize these nuances and build adaptable operating models accordingly are the ones that consistently capture my attention as an investor.

There’s a growing focus on real-time analytics and decision automation. How prepared are businesses in Southeast Asia for this shift, and what infrastructure is often missing?

Southeast Asia’s readiness for real-time analytics reminds me of the famous William Gibson quote: “The future is already here – it’s just not evenly distributed.” In Singapore’s financial district, you’ll find institutions with analytics capabilities rivaling anything in New York or London. Travel a few hundred kilometers in any direction, and you’ll encounter businesses struggling with fundamental data integration challenges.

This uneven distribution creates a dynamic where organizations skip entire generations of incremental improvements to adopt cutting-edge solutions. A logistics company in Vietnam might move directly from basic spreadsheet analytics to IoT-enabled real-time fleet optimization, bypassing the intermediate steps many Western companies trudged through.

The infrastructure gaps holding back real-time analytics in the region fall into three categories. The first is technical – inconsistent cloud access, limited API standardization, and aging enterprise systems that weren’t designed for continuous data flows. The second is human – a shortage of professionals with both the technical skills to implement real-time systems and the business acumen to apply them effectively. The third is organizational – decision processes designed for monthly or quarterly rhythms that can’t capitalize on minute-by-minute insights.

What’s often overlooked in discussions about Southeast Asia’s real-time readiness is the impact of sovereignty considerations. Governments across the region are increasingly implementing data localization requirements, which can complicate the deployment of global real-time solutions. A financial institution might have the technical capability for instantaneous cross-border analytics but face regulatory barriers that introduce friction.

Despite these challenges, I’ve witnessed remarkable innovation in pragmatic implementations. The Ascott Group, for instance, sponsored a hackathon at last year’s SWITCH festival in Singapore, where one of the winning teams developed an IoT-based system promoting sustainable habits among Lyf guests. The solution displayed guests’ real-time carbon footprint impact on the property and automatically released rewards when sustainability tasks were completed. What made this approach particularly effective was its hybrid nature – certain critical sustainability metrics were monitored in real-time, while less urgent data followed traditional batch processing. This pragmatic prioritization matched their actual decision-making needs rather than pursuing real-time capabilities indiscriminately.

The businesses succeeding in this transition aren’t necessarily those with the biggest technology budgets. They’re the ones aligning their analytics velocity with their decision velocity – ensuring that the pace of information matches the pace of action. That alignment, more than any specific technology, determines readiness for the real-time future.

With your work in fintech and data intelligence, how do you see regulatory compliance intersecting with innovation, especially when dealing with consumer data or financial insights?

The relationship between regulatory compliance and innovation in Southeast Asia is nothing like the adversarial dynamic often portrayed in Silicon Valley narratives. It reminds me of the concept of “scaffolding” in developmental psychology – temporary support structures that enable growth rather than constrain it.

Through my journey in pursuing the Chartered Fintech Professional designation and my experience at Attribute Data, I’ve observed how thoughtfully designed regulatory frameworks actually create the conditions for sustainable innovation. The Monetary Authority of Singapore’s FinTech Regulatory Sandbox, for instance, functions less as a restrictive boundary and more as a protected space for experimentation, allowing fintech innovators to test concepts with real consumers under controlled conditions.

What makes Southeast Asia particularly interesting is how regulatory approaches to data and financial innovation vary dramatically across borders. Singapore’s “regulation-as-enabler” philosophy contrasts with more cautious approaches in other markets, creating a natural laboratory for different models of innovation governance. Companies operating regionally must navigate these variations while maintaining consistent ethical standards – a challenge that often spurs rather than inhibits creative thinking.

Consider how this plays out with consumer financial data. When working with financial institutions across the region, we’ve taken on “Embedded Governance” methodologies – integrating regulatory requirements, data privacy concerns, and bias mitigation directly into data architectures from the foundation. This approach transforms compliance from a cost center to a competitive advantage through enhanced customer trust. We pioneered bringing GDPR-compliant technologies to Asian markets, coming up with GDPR-readiness checklists for our global clients before GDPR and local regulations like Singapore’s PDPA were even in force.

Singapore is breaking new ground globally through initiatives like the Data Trust Mark and the AI Verify Foundation. The latter has developed the world’s first comprehensive AI governance testing framework and software toolkit that helps industries demonstrate transparency in their AI systems. These frameworks provide practical tools for responsible innovation rather than merely imposing restrictions.

The most successful innovators I’ve encountered don’t ask “how can we work around regulations?” but rather “how can we harness regulatory frameworks to build more trustworthy products?” This mindset shift is crucial in Southeast Asian markets where digital financial services are still building public confidence.

What’s often overlooked in discussions about regulation and innovation is the role of cultural attitudes toward data privacy. Western frameworks like GDPR reflect specific cultural assumptions about individual data ownership that don’t always align with perspectives in certain Asian contexts where collective benefits might be weighted differently. Navigating these nuances requires regulatory humility – recognizing that there’s no universal “right” balance between innovation and protection.

The organizations that thrive at this intersection are those that engage proactively with regulators, viewing them as stakeholders in innovation rather than obstacles to it. This collaborative approach, more than any particular technology or business model, distinguishes the most promising fintech or data intensive startup developments I’m witnessing across the region.

As someone at the intersection of data, AI, and venture capital, what are you most optimistic about in the next 12 to 18 months—and where do you think people are getting it wrong?

I’m looking at the evolution of “contextual AI” in Southeast Asia—moving beyond generic foundation models to create solutions that truly understand local nuances. While the past 12 months saw an explosion of new LLMs with strong corporate backing entering the market, the next 12-18 months will be defined by their practical application to regional challenges through several key developments.

First, we’ll see the rise of what some industry insiders call “invisible AI”—collaborative systems working behind the scenes to streamline operations across sectors. Unlike flashy consumer applications, these tools will focus on automating compliance processes, enhancing transaction monitoring, and enabling predictive maintenance in manufacturing. The value proposition isn’t the AI itself but the operational efficiency it unlocks.

Second, I expect rapid advancement in vertical-specific AI solutions. Rather than generic models attempting to solve everything, we’re witnessing the emergence of specialized systems for healthcare diagnostics, legal document analysis, and construction risk assessment—particularly valuable in Southeast Asia’s diverse regulatory environments. These vertical SaaS solutions leverage domain-specific data to deliver precision that general models cannot.

Small Language Models (SLMs) and agentic AI are particularly promising for our regional context. While massive models grab headlines, compact task-specific AI offers more cost-effective deployment for customer service, fraud detection, and workflow automation. This is crucial for Southeast Asian markets with varying levels of digital infrastructure. Additionally, agentic AI systems that can autonomously execute multi-step tasks—like loan approvals or supply chain optimizations—are becoming increasingly sophisticated.

During a trip to China with SMU IIE as part of the LKYGBPC outreach team, I visited i2Cool, a Shenzhen-based startup developing electricity-free cooling technology. Their approach highlights another trend I’m watching closely: climate-tech AI applications. From optimizing energy consumption to predicting equipment failures in renewable energy systems, AI is becoming essential for sustainability initiatives—a critical concern across Southeast Asia’s rapidly developing urban centers.

On the hardware front, custom silicon developments will significantly impact AI adoption. The partnerships forming between startups and chipmakers to optimize AI workloads for specific industries will democratize access to computational resources that were previously prohibitively expensive, enabling more Southeast Asian companies to deploy sophisticated AI solutions.

Where I think many are getting it wrong is in what behavioral economists call “availability bias”—overweighting information that’s readily available. There’s excessive focus on large, general-purpose AI models because they generate impressive demos, while the transformative potential in specialized applications receives less attention despite potentially greater impact. The most valuable innovations won’t necessarily be the most visible ones.

The most common misconception I encounter is that Southeast Asia is simply following innovation patterns established elsewhere with a time delay. The reality is far more interesting—we’re developing hybrid approaches that combine elements from different global frameworks with distinctive local innovations that address unique regional challenges. What ultimately makes me optimistic is seeing AI development increasingly aligned with fundamental societal challenges—from financial inclusion to climate resilience.

I’d like to take this opportunity to call out to deep tech startups solving urban solutions and sustainability challenges to consider joining the Lee Kuan Yew Global Business Plan Competition. The deadline closes on April 30, 2025, and it’s a tremendous platform for startups with transformative ideas to gain visibility and support. When technology development aligns with these deeper purposes, both innovation and human well-being advance together.