Singaporean companies report losing about USD half a million (S$600,000) each year due to poor software quality. Often, this is due to technical debt, where companies take shortcuts, such as skipping tests, applying quick fixes, or rushing software releases.
These issues rarely appear in financial reports but create fragile systems that disrupt operations and damage customer trust.
“Over a third of global teams cite unresolved technical debt as the biggest barrier to software quality, with many admitting to pushing out untested code,” Tricentis’ senior vice president, APAC & Japan, Damien Wong, told iTnews Asia.
Legacy systems, including monolithic applications that are hard to update or platforms running on ageing infrastructure that cannot integrate with modern workflows, remain another major risk, Wong added.
Across APAC, the risk is growing as companies race to modernise. As per Forrester, software and IT services will drive 66 percent of global tech spending in 2025, much of it to modernise legacy systems.
The rush to move faster again leads to shortcuts, leaving businesses caught between the pressure for speed and the risk of fragile systems.
Monitor signals before they escalate into a crisis
Wong mentions the need to treat technical debt as a measurable risk rather than a vague concept, and there is a need to monitor it.
One key practice is test gap analysis, where teams identify areas of new code that remain untested, as such gaps represent high-risk zones and can lead to costly failures.
“We put this into practice using SeaLights Quality Intelligence platform, which automatically analyses code changes and highlights any gaps in coverage testing,” said Wong.
Secondly, one needs to track test coverage trends.
Instead of chasing 100 percent coverage, it focuses on testing the most critical and frequently changing parts of the code.
A sudden drop in meaningful coverage or a rise in low-value tests immediately signals potential risk.
Technical debt reveals itself in test automation maintenance.
Organisations that rely on script-based automation struggle as systems expand and integrate.
A single change in one application can ripple across others, making test assets difficult to maintain.
- Damien Wong, Senior Vice President, APAC & Japan, Tricentis
To counter this, Wong said organisations need to move towards AI-driven model-based test automation, which automatically updates test assets and ensures end-to-end software quality remains consistent.
He added that in Singapore alone, over 80 percent of organisations expect significant productivity gains from applying generative AI to software quality efforts.
Why slow is the new downtime
Businesses need to adopt a continuous quality engineering strategy.
Wong mentioned that the approach builds reliability into products from the start, catching defects, inefficiencies, and architectural issues early, before they grow into costly problems or add to technical debt.
To succeed, Wong firstly stresses automation as a key investment.
AI-driven model-based testing reduces manual work, speeds up response to change, and spots risks earlier in development.
In addition, Wong said teams need to use AI-driven model-based testing to reduce manual work, respond quickly to change, and spot risks earlier in development.
Performance testing strengthens protection by checking APIs and services early, uncovering latency and scalability issues before they affect users.
In today’s environment, where “slow is the new downtime,” anticipating performance risks becomes crucial.
Second, Wong added that it is equally important to have a culture of shared ownership of quality.
When developers, testers, and product managers work together, teams build resilient systems and avoid late-stage fixes.
This helps organisations deliver faster while maintaining reliability.
AI can tackle technical debt but demands human oversight
AI can play a central role in reducing technical debt across the software lifecycle. First, Wong said AI can enable smoother integration of legacy systems.
Enterprises still rely on outdated architecture, and AI-powered testing tools can help abstract business logic away from systems, simplifying validation and making legacy modernisation achievable, said Wong.
Second, AI can help prevent future debt.
Wong said model-based testing allows teams to generate test frameworks directly from design prototypes.
That means quality and testability are built in from the earliest stages of development, he added.
Third, AI accelerates test maintenance.
One of the biggest challenges in testing is keeping pace with constant application changes.
Citing a case study, he mentioned how Worldpay leveraged Tricentis Tosca to automate testing across a complex data warehouse environment.
This resulted in a 90 percent reduction in maintenance effort while maintaining comprehensive compliance and coverage.
Finally, it democratises testing.
Wong said no-code AI tools allow non-technical stakeholders to review and validate requirements, breaking down silos between business and engineering.
This ensures that what's being built matches what's needed before it becomes expensive to fix.
On the other side, AI can also introduce new risks, including hallucinated outputs or compatibility issues.
To counter this, Wong explains that AI tools should act as co-pilots, not replacements for human input.
Using testing-specific metadata and techniques such as retrieval-augmented generation (RAG) keeps outputs aligned with real business goals.
With proper human oversight and domain expertise, AI can be harnessed to reduce technical debt and make teams agile.
The real challenge, Wong said, is translating technical debt into boardroom language.
Software quality is a business value
CIOs need to shift their view of software quality from being a cost centre to a driver of business value.
Wong noted that mature DevOps teams utilising AI in testing are almost 30 percent more likely to rate themselves as highly effective, with testing emerging as the highest-return area of AI investment.
By investing in software quality testing technology and practices, organisations can see tangible business outcomes, including a reduction in production errors and time-to-release, said Wong.
Organisations that invest in modern testing technologies and practices cut production errors, release software faster, and strengthen revenue, customer loyalty, and retention, he added.
.jpg&h=271&w=480&c=1&s=1)
