SINGAPORE, March 12, 2026 – Across Southeast Asia, artificial intelligence is moving rapidly from experimentation to enterprise deployment. Yet many organisations are discovering that the biggest obstacle to scaling AI is not the sophistication of the models, but the condition of their data.
Joseph Bosco, Partner Manager for Asia Pacific and Japan at Databricks, believes this realisation is fundamentally reshaping how enterprises approach AI transformation in the region. As companies shift from proof-of-concept projects to production-scale AI systems, partners are increasingly being called upon to modernise fragmented data estates, build unified platforms and establish governance frameworks that allow AI to operate safely and reliably.
“The shift I’m seeing is very simple. Most customers are realising that AI is only as good as their data. Early on, everyone wanted a shiny large language model demo,” Bosco explains. “Now the conversation is about how to build reliable, governed data foundations that can support AI over the long term.”
Bosco believes that this shift is redefining the role of technology partners across Southeast Asia, particularly as enterprises begin to realise that AI outcomes are fundamentally tied to the quality and accessibility of their data.
From AI Experiments to Data-Driven Platforms
Across Southeast Asia, organisations are experimenting with AI applications ranging from fraud detection and customer analytics to supply chain optimisation. However, many enterprises still operate with fragmented data environments spread across legacy infrastructure, multiple clouds and SaaS platforms.
“If your data is fragmented, your AI will be fragmented. It’s that direct,” Bosco says.
This is where partners increasingly play a strategic role. Rather than focusing solely on implementing tools, partners are helping enterprises consolidate data into unified platforms, modernise legacy pipelines and introduce governance frameworks that allow AI models to operate safely at scale.
Bosco points to Databricks’ lakehouse architecture as an example of how enterprises are converging structured, semi-structured and real-time data into a single environment, enabling data engineering, analytics and machine learning teams to work on the same foundation.
Tools such as Databricks’ AI assistant Genie are also making this transformation more visible to business leaders.
“With Genie, partners can sit down with a business user and allow them to interact with their own data through natural language,” Bosco explains. “That experience quickly highlights which pipelines, governance policies and data sources need to be fixed to make AI production-ready.”
A Highly Diverse Regional Data Landscape
Unlike North America or Europe, Southeast Asia presents a highly varied data landscape where enterprises operate at very different stages of digital maturity.
Bosco notes that while markets such as Singapore may already be deploying sophisticated data mesh architectures, organisations in neighbouring countries may still be grappling with data stored in legacy systems or spreadsheets.
“Southeast Asia is not one data market. It is many different data economies sitting next to each other,” he says.
This diversity makes local partners particularly important. Regional system integrators and consulting firms often possess deep knowledge of where enterprise data resides, how it is governed and which legacy systems remain embedded in business processes. That local expertise, Bosco says, is often the difference between theoretical AI strategies and real-world deployment.
“Partners who understand where the data actually lives in a bank, telco or government agency become critical in turning platforms like Databricks into vehicles that leverage local regulation, language and culture,” he adds.
Turning Data Platforms into Business Outcomes
Across Southeast Asia, partners are already helping enterprises translate AI ambitions into tangible business results.
In financial services, for example, banks are using unified data platforms to combine transactional records, customer interactions and risk datasets. Machine learning models built on this shared foundation are enabling more effective fraud detection, anti-money laundering monitoring and customer engagement strategies.
“The outcome isn’t simply that a bank has a model,” Bosco explains. “It’s lower fraud losses, faster case handling and higher cross-sell opportunities, all traceable back to well-governed data.”
In digital-native companies, partners are consolidating clickstream data, behavioural analytics and transaction histories to power real-time recommendation engines and customer personalisation systems.
Retail and consumer goods companies are similarly integrating point-of-sale data, inventory information and supply chain analytics to improve demand forecasting and operational efficiency.
Once these data foundations are established, tools such as Genie enable business teams to explore performance trends and generate new ideas directly from trusted enterprise data.
From AI Pilots to Scalable Services
For technology partners themselves, the AI opportunity is also evolving rapidly.
Bosco observes that many partners initially enter the AI market through small proof-of-concept projects and experimental demonstrations. However, the real opportunity lies in building repeatable services that enterprises can deploy at scale.
“The turning point is when partners realise their primary product is not a model. It is a repeatable way of turning raw data into reliable applications,” he says.
This typically involves building standardised data pipelines, machine learning operations frameworks and governance layers that support multiple use cases. Partners can then package these capabilities into industry-specific solutions such as telecom churn prediction platforms or generative AI assistants for banking operations.
Over time, these offerings evolve into dedicated AI practices that include platform engineers, data engineers, machine learning specialists and governance experts.
Governance and Trust in the GenAI Era
As generative AI adoption accelerates across the region, enterprises are also becoming increasingly concerned about governance and regulatory compliance.
Bosco emphasises that trust in AI ultimately depends on strong data governance frameworks.
Within the Databricks ecosystem, governance is managed through Unity Catalog, which provides a centralised layer for organising data and AI assets, controlling access and maintaining audit trails.
“Partners play a major role in translating regulatory policies into practical data controls,” Bosco explains. “They help organisations decide which datasets can feed generative AI models, how personal data is protected and how AI systems are monitored over time.”
In regulated sectors such as banking and telecommunications, partners often design architectures where AI models operate within tightly controlled environments, with full traceability for every query and output.
The Future of the AI Partner Ecosystem
Looking ahead, Bosco expects the AI partner ecosystem in Southeast Asia to expand significantly over the next three to five years as enterprises move beyond experimentation.
Partners that master data platform implementation, governance frameworks and AI operationalisation will increasingly become long-term strategic advisors to organisations across sectors.
“Partners who can take fragmented data, govern it properly and build AI capabilities on top of it will effectively help design how organisations make decisions for the next two decades,” Bosco says.
He also expects the emergence of specialised AI partners developing industry-specific data products, reusable generative AI architectures and decision-support tools built directly on enterprise data platforms.
In Southeast Asia’s rapidly evolving digital economies, Bosco believes these partner ecosystems will be central to unlocking the full value of AI.
“The future of AI in this region will not be driven by technology platforms alone,” he says. “It will be driven by the ecosystems that help enterprises turn data into real operational advantage.”