Businesses race to adopt AI, but ROI remains an open question

Businesses across Asia-Pacific are stepping up investment in artificial intelligence (AI), yet many organizations admit they have not fully assessed the effectiveness of these investments, largely due to pressure to keep pace with technology trends and avoid falling behind.

These findings come from a newly released survey conducted by market intelligence firm International Data Corporation (IDC) on behalf of Expereo, a Netherlands-based managed network services provider. The survey covered 800 technology leaders at multinational enterprises with more than 500 employees across the United States, Europe and Asia-Pacific.

According to the survey, about 37% of organizations in Asia-Pacific said they had invested heavily in AI but had done little to assess outcomes. Singapore is seen as one of the region’s AI deployment hotspots, with more than one-third of organizations in the city-state saying they had made significant AI investments without fully evaluating the technology’s real-world impact.

The findings reflect an intense race to adopt AI across the business community, despite persistent difficulties in proving return on investment (ROI) and building infrastructure capable of supporting large-scale technology deployment.

“Every enterprise we speak with is investing in AI, but the data shows a clear gap opening up between AI ambition and outcomes. Often, that gap starts with network infrastructure. AI can only deliver on its promise when the infrastructure supporting it is built to meet that demand.

Without robust, scalable, and cloud-optimized networks, even the most heavily funded AI programs will struggle to deliver returns on investment. Building the right network is no longer a decision for the IT department alone; it is one of the most important boardroom conversations taking place today to turn AI ambition into reality,” said Ben Elms, CEO of Expereo.

The report found that only 40% of organizations in Asia-Pacific said their AI projects had met or exceeded expectations, a figure higher than the global average. The biggest barriers cited by businesses include poor training data quality, higher-than-expected implementation costs and underperformance of AI models.

Globally, only 19% of survey respondents said their AI initiatives had exceeded expectations, while just 5% said results had significantly surpassed initial expectations.

Alongside investment effectiveness, long-term costs are becoming a growing concern. The survey found that 41% of technology leaders in Asia-Pacific are concerned about losing control over AI-related costs and returns. In Malaysia, the figure rises to 54%.

Even so, confidence in AI’s potential remains strong. As many as 61% of organizations in the region said AI and machine learning would remain among their top technology investment priorities over the next 12 months.

The survey also highlights significant technology infrastructure constraints. Only 9% of organizations in the Asia-Pacific believe their current networks are ready to support new AI, cloud computing, and digital transformation initiatives. Meanwhile, 37% said their networks would need to be upgraded or replaced in the period ahead.

According to Eric Wong, President of Asia-Pacific at Expereo, businesses are increasingly scrutinizing whether their networks and technology systems can support large-scale AI applications.

“Organizations that address these foundational issues early will see better results and faster operational impact,” Wong said.

In Vietnam, the challenges facing businesses in AI adoption are also substantial. Beyond the pressure to prove returns and upgrade technology infrastructure, many companies face constraints in human resources and data quality, two foundational elements of any AI strategy.

According to a 2025 survey on AI adoption in Vietnam conducted by AWS in partnership with Strand Partners, 55% of Vietnamese businesses identified a lack of digital skills as the biggest barrier to digital transformation and AI adoption. A shortage of specialists in data, machine learning, and AI has made it difficult for many businesses to deploy and operate projects at scale.

Talent is not the only constraint. Many organizations also face major data challenges. Although data is often considered the “fuel” of AI, data at many Vietnamese enterprises remains fragmented, inconsistent, unstandardized, or scattered across separate systems. This significantly reduces the output quality of AI models.

In addition, the upfront cost of AI investment remains a major challenge, including technology infrastructure, software licensing, data storage and processing costs, and the recruitment of highly skilled specialists. At the same time, measuring the return on investment of AI projects is often complex and requires time to prove effectiveness, making many businesses more cautious in implementation.

Even so, experts believe AI is no longer an experimental option but is becoming an essential requirement for maintaining competitiveness in the digital era. Instead of following trends or spreading investment too thinly, businesses need to approach AI strategically, starting with the identification of specific business problems and the development of capabilities in operations, data analytics, and customer engagement.

The success of AI depends not only on technology but also on long-term commitment from leadership and investment in people, data, and infrastructure. Businesses that address these foundational issues early will have a stronger chance of turning AI’s potential into real competitive advantage in the years ahead.