AI is scaling fast, but so are its mistakes
APAC’s AI Problem Isn’t the model. It’s the data. For many, AI has moved well beyond the experimental phase and into core infrastructure.
But acceleration is exposing a problem that was always there. AI outputs are increasingly irrelevant, mistimed, or just wrong. Customers receive offers for products they have already bought. Recommendations feel disconnected from any real understanding of who they are. Interactions designed to feel personal instead feel intrusive.
These are not model failures. They are data failures, and AI is making them louder, faster, and more visible.
AI doesn’t fix bad data, it amplifies it
Organizations employ AI to drive growth and efficiency at scale. What often gets overlooked is that the quality of data feeding those systems scales, too. According to Gartner, poor data quality costs organisations an average of $12.9 million per year, and that figure predates the AI era, when the consequences of bad data moved far more slowly.
Today, a fragmented or delayed data feed does produce a bad customer or user experience at machine speed, replicated across existing touchpoints simultaneously.
A banking app pushing irrelevant offers based on six-month-old financial behavior. A retail platform recommending items that a customer returned weeks ago. A superapp serving content in the wrong language to the wrong segment. These are data failures, surfacing faster and at greater scale.
APAC’s data infrastructure wasn’t built for this
Despite investments in AI, many are still operating on data infrastructure that was never designed for real-time decisioning. Customer data sits fragmented across CRM platforms, mobile apps, and websites, with no unified view. Processing remains heavily batch-based, meaning insights arrive too late to be useful, and many organizations still depend on third-party or inferred data that lacks both accuracy and transparency.
This is a particular challenge in APAC, where the regulatory and market landscape is evolving fast. Governments across the region are tightening data privacy requirements, and organizations must navigate cross-border data flows while staying compliant, often across markets with very different requirements. At the same time, customer expectations are rising. Consumers want experiences that are personalized and trustworthy. They will accept relevance, but not at the cost of feeling surveilled. That is a difficult balance to strike when your data foundation is fragmented, delayed, and opaque.
AI now operates in milliseconds, yet for many organizations, the data it depends on is often days or weeks behind. That gap is rapidly becoming a competitive liability.
Infrastructure before personalization
For years, personalization has been framed as the goal. In the age of AI, it is the outcome. Organizations that will lead the next phase of adoption are those on foundations that make those models work. Success will be those who shift their focus from surface-level use cases to the underlying data infrastructure that powers them.
This means building a real-time unified view of every customer touchpoint to deliver a competitive advantage; having the right data, at the right moment, governed in the right way.
The race to AI adoption in APAC is real, and the pressure to move fast is justified. But speed without data integrity does not produce a competitive advantage — it produces confident mistakes at scale.
The organizations that will define the next era of customer experience are not necessarily those that deployed AI first. They are those who asked the harder question first: Is our data actually ready for this?
That is where the real work begins.
Marie Dalton is Vice President of Marketing for APJ at Tealium, where she leads regional strategy across demand generation, digital, field, and partner marketing. With over 20 years’ experience in B2B SaaS, she specializes in building high-performing marketing organizations that drive pipeline growth and align closely with sales. Based in Sydney, Marie is focused on helping enterprises unlock the value of real-time customer data to power AI and better customer experiences.
TNGlobal INSIDER publishes contributions relevant to entrepreneurship and innovation. You may submit your own original or published contributions subject to editorial discretion.
Featured image: Yamu_Jay on Pixabay