Enterprises seem to rush to implement AI only to realise that the technology is ineffective without high-quality, well-organised internal information in a structured, valuable form. As a result, Gartner predicts that organisations will abandon 60% of AI projects unsupported by AI-ready data in 2026.
The Kazunori Fukuda, managing director of Sansan Global for Singapore & Thailand, discusses the gap between AI investments and the value they bring and valuable considerations to help bridge it.
AI ROI gap
According to Fukuda, investing in AI but struggling to see the expected returns in productivity or efficiency creates an AI ROI gap, especially when AI tools are not fully integrated into business processes or when data quality is poor.
"For example, in Southeast Asia, many businesses face fragmented data infrastructures and limited digitalisation. As a result, AI tools may not be able to effectively process data, leading to high costs without delivering tangible business value," he said.
He argued that investing in AI-driven solutions while still dealing with inconsistent data across departments will not streamline processes. Similarly, investing in generative AI but failing to align the data infrastructure with the AI tools' capabilities can add to the frustration because the automation does not deliver the expected time and cost savings.
"Instances like these are where the AI ROI gap becomes clear. AI wasn't the problem, but the data wasn't ready to support it," he said.
AI wasn't the problem, but the data wasn't ready to support it. Kazunori Fukuda
Causes of the AI ROI gap
This gap does not stem from a single failure, but from several structural mismatches. Fukuda says the AI ROI gap often arises from a mismatch between the AI technologies businesses adopt and the actual infrastructure or processes they have in place.
"Businesses invest in advanced AI tools only to face challenges when trying to integrate them into existing workflows," he explained.
In Southeast Asia, for example, it is quite challenging to leverage AI effectively because many companies still rely on fragmented data systems that aren't fully digitised.
Japan, on the other hand, is geared towards more centralised data management systems. However, concerns around data privacy and regulatory risks slow AI adoption.
Fukuda added: "Moreover, insufficient integration of AI into existing systems and workflows, as well as a lack of skilled personnel, exacerbate the gap."
Measuring AI ROI
Because of these challenges, how companies define and measure ROI becomes critical. Companies tend to measure AI ROI only by the productivity it delivers to the workforce.
However, there are other metrics beyond productivity, such as operational efficiency, decision-making quality, and customer satisfaction, according to Fukuda.
"AI's ability to automate repetitive tasks and deliver actionable insights can significantly improve decision-making processes. For example, AI can help businesses in APAC markets like Southeast Asia make faster, more informed decisions by analysing large datasets more quickly than manual processes ever could. AI also plays a role in improving customer experiences by delivering personalised services, faster response times, and more accurate insights," he said.
Fukuda says that their company measures the impact of AI not only by efficiency gains but also through the quality of decisions it supports. Moreover, he cites improvement of customer satisfaction as a key metric.
"Our AI-powered solutions, which streamline tasks like business card digitisation and data management, have enabled teams to spend more time on value-added activities, which directly contribute to a better customer experience and reduce the cost of digitalisation," he said.
AI-enabled information infrastructure in SEA
The strength of Southeast Asia as a testbed lies in its variety, as AI solutions can be adapted and refined in response to specific local needs, allowing businesses to scale these solutions across the region once they've proven effective." Kazunori Fukuda
These ROI challenges play out differently across regions, particularly in Southeast Asia.
For Fukuda, the region's rapid digital transformation and diverse business environments make it "an exciting region for AI innovation".
"From experiences working across markets like Singapore, Thailand, Vietnam, and the Philippines, I've seen firsthand how AI can solve significant inefficiencies," Fukuda said.
He said that industries such as logistics, manufacturing, and construction have significant opportunities to streamline operations with AI, as many still rely heavily on manual, paper-based processes.
He considers its diversity a double-edged sword because it also has challenges, mainly when countries differ in digital maturity, infrastructure, and regulatory frameworks.
Businesses in Thailand, for example, struggle to leverage AI effectively due to fragmented data concerns.
However, AI-driven solutions can significantly enhance business performance by automating tasks.
The Sansan executive said: "The strength of Southeast Asia as a testbed lies in its variety, as AI solutions can be adapted and refined in response to specific local needs, allowing businesses to scale these solutions across the region once they've proven effective."
Based on his experience, Fukuda said that distinct cultural and regulatory factors shape data challenges in Southeast Asia and Japan.
"Businesses in markets like Thailand and the Philippines often deal with inconsistent digital adoption, relying on a mix of paper-based and digital systems. This creates barriers for leveraging AI or automating processes effectively," he said.
On the other hand, Japan has a more centralised data management system with strong data governance and protection.
"However, the cultural mindset in Japan is more risk-averse, which can slow the adoption of new technologies, particularly AI. Japanese companies tend to prioritise privacy and continuity over disruption, making them cautious about the pace of technological change," Fukuda explained.
He advises CIOs to focus not only on technical capabilities but also on understanding regional nuances to build AI solutions successfully.
"To avoid the industry's high project abandonment rates, leaders must move beyond a tool-first mindset and prioritise data-first business infrastructure that aligns with the specific cultural and operational realities of each market," he said.
Success in AI initiatives
With organisations predicted to abandon 60% of AI projects unsupported by AI-ready data in 2026, Fukuda said that tech leaders must ensure that AI is aligned with clear business goals and deeply integrated into core operations.
He posits that AI adoption success involves ensuring AI fits within existing workflows, which requires a solid data infrastructure, continuous employee upskilling, and constant evaluation of the technology's performance.
"Many businesses face challenges when AI is introduced without a clear strategy or defined objectives. This can result in the technology not integrating well with current systems, leading to wasted resources and unmet expectations. AI is not a one-time solution; it's an ongoing journey," he added.
A good practice in their company is fostering collaboration across departments to refine and adapt AI tools continually. It is also vital to let AI evolve alongside the company's needs, and leaders should incorporate it into the business strategy.
Bridging the AI ROI Gap?
Taken together, these lessons point to one central requirement. To bridge the AI ROI gap, Fukuda urges CIOs to prioritise building a solid foundation of structured, high-quality data before deploying AI.
"The AI ROI gap often widens when businesses deploy AI solutions without having their data properly organised. In Southeast Asia, many businesses are still grappling with siloed data and paper-based systems, which can limit AI's ability to deliver the expected results," he explained.
One of the best practices they follow at Sansan is ensuring that data is clean, organised, and easily accessible, enabling seamless AI integration into their solutions. This practice provides the company with better insights, reduces costs, and improves overall customer experience.
"CIOs should be used to aligning AI projects with clear business objectives and ensuring that data quality is prioritised from the outset. AI can't deliver results if the data isn't ready to support it," he concluded.
Successful AI initiatives are no accident. With the right tools, mindset and practices, businesses can maximise the value of their AI investments.
