Who doesn’t like shortcuts?
Definitely not data professionals. Poor data decisions may work in the short term, but the costs come later, and they compound at scale.
Steven Seah, area vice president, South Asia, Informatica from Salesforce, said: “Every shortcut that is taken with data, from undocumented fields and duplicated records to integrations with little or no governance behind them, doesn’t disappear. It compounds over time.”
These costs, commonly known as “data debt,” are already costing organisations an estimated $3.1 trillion annually.
“Beyond financial loss, unaddressed data debt can also hamper innovation, slow decision-making, increase risk, and drive up the cost of change,” Seah added.
AI and data debt
AI cannot deliver value without trusted data. Steven Seah
The challenge becomes even more evident with AI adoption accelerating. Preventing data debt, ensuring data reliability, and enforcing governance have become imperative for organisations.
“AI cannot deliver value without trusted data,” Seah said. “If your underlying data is inconsistent or unreliable, it becomes an operational and business risk that compounds at scale.”
In Southeast Asia, Seah observes highly fragmented data landscapes, largely because systems are built in silos and integrations are layered on over time rather than designed holistically. This results in unclear ownership, duplicated records, and declining trust in data.
“AI adoption is accelerating faster than governance can keep up, and this is where data debt starts to compound. More than three-quarters (76%) of data leaders say their company’s visibility and AI governance have not completely kept pace with employees’ use of AI, showing how quickly AI deployment is outpacing the governance needed to support it,” Seah said.
Inconsistent, unreliable, or poorly governed data makes AI ineffective. The competition is shifting from who can move fastest in AI implementation to who can build the strongest foundation of trusted data.
In this way, Seah believes that speed need not come at the cost of reliability.
As AI seeps into business operations, the costs of poor data will eventually surface across the organisation.
The invisible costs
The negative effects of poor data are not always immediately measured or reported.
Instead, they surface through missed customer insights, stalled AI pilots, and the absence of trusted numbers to support business decisions.
“It surfaces as operational inefficiencies, missed revenue, and eroding customer trust. These do not show up on an organisation’s P&L, and they risk being normalised instead of addressed,” Seah said.
Data reflects this growing concern. Nearly half of data leaders (49%) said they have reached incorrect conclusions because of gaps in business context, while two in five (42%) report low confidence in AI outputs when data context is missing. Gartner estimates that poor data quality costs organisations an average of $13 million a year. The stakes even become higher with the rise of agentic AI.
“A human might pause and question a data point that looks wrong, but an agent won’t. Agents act autonomously at scale with full confidence. The real invisible cost of data debt is not what it costs you today, but the liability you inherit the moment AI agents are acting on your behalf. That makes every decision an expensive guess,” Seah said.
Data governance
Poor data creates invisible costs. Governance is what helps organisations prevent them.
For Seah, data governance is not just about compliance but an enabler of trust and transparency.
“Data governance serves as the foundational layer that gives organisations visibility into where data comes from, how it is used, and whether it meets quality and compliance standards, empowering organisations to confidently leverage data as a trusted asset.”
Putting governance into practice requires the right technology. This is where modern data platforms come in. According to Seah, they help unify fragmented data, resolve inconsistencies, and create a consistent view across systems.
“This is what we refer to as trusted context. It ensures data is accurate, governed, and meaningful so AI can deliver reliable outcomes. Without that foundation, AI initiatives don’t just struggle to scale. They can introduce more risk than value.”
A solid data foundation enables governance to help teams move faster, not slow them down.
“No one is stuck reconciling numbers or chasing down where a field lives, teams are empowered to take action. And in an agentic world, that role becomes even more important. The platform isn’t just storing and cleaning data; it’s providing the layer that gives AI agents the trusted context they need to act.”
Addressing the invisible costs of data debt
Your AI systems are only as good as the people who interpret and trust their insights. Steven Seah
Beyond technology, organisations also need to employ the right strategies to address data debt.
“CIOs need to answer a fundamental question: Do we actually trust the data underneath our AI? It is easy to let AI ambitions outrun the data foundation, and that’s when AI projects and pilots fail.”
Seah believes Southeast Asia has a unique opportunity to build AI-ready foundations from a cleaner slate than markets burdened by decades of legacy infrastructure.
Seah urges organizations to audit their data debt and understand their data landscape early, so they can prioritize actions and mitigate risks effectively.
Once organisations understand their data more deeply, Seah said the next step is to embed trust in the data foundation early on.
“Connect systems, improve data quality, and put robust governance in place. Leading organisations treat data as a strategic asset because trusted AI starts with trusted data,and that’s where the real competitive advantage emerges.”
Seah also recommends investing in data and AI literacy across the entire workforce, not just within IT or data teams.
“Your AI systems are only as good as the people who interpret and trust their insights,” Seah concluded.
No shortcuts
Some shortcuts may lead to the same destination, but never the same effect.
When it comes to data, organisations must be willing to invest in building the right foundations first to ensure it leads to broader advantages in the long term.
