Could the high perceived investment/effort factor be an impediment? Or is the unimpressive ROI due to mindset or implementation mistakes?

Is the Asia Pacific region really ready to make the most of AI?

According to some surveys, about 70% of organizations polled in the region recognized the potential of AI, but only 17% were at the highest stages of AI maturity, compared to less to half of manufacturing and supply chain industries considered as lagging in AI adoption.

Given the pandemic-driven need to digitalize, and the gap between AI awareness and adoption maturity is surprising. Could this low level deliberate or planned by management in the affected industries?

According to Leon Lim, CEO & Founder, Groundup.AI, there are some dynamics behind the numbers. He shares his market insights with DigiconAsia in an exclusive interview.

DigiconAsia: How does organization size affect the rate of adopting AI and achieving maturity?

Leon Lim (LL): The challenges in adoption and maturity come in the following forms, linked to corporate size, infrastructure and mindsets:

  • inadequate infrastructure
  • lack of talent with the IT/AI/Data science skills sets
  • financial considerations
  • the fact that over more than 400m people in APAC alone do not have access to basic internet services. This is a huge stumbling block as many AI technologies require such basic infrastructure.

Roughly only 16% of small- and medium- sized enterprises in the region employ advanced data-driven technologies, with half citing cost as a major factor. Meanwhile, even big companies face challenges that have hindered AI adoption, with difficulties integrating new solutions into their existing legacy systems being one of the top reasons. This may lead to more reluctance and longer decision-making processes whenever they want to implement something such as AI. 

To address the slow adoption, it is important for firms in the APAC region to develop core IT infrastructures and data ecosystems, while building a talent pool to reap the benefits of AI in the long run. 

DigiconAsia: Could some organization lagging in AI be starting from a low base and feel they need to get to grips with baseline digital transformation first before they can cope with the additional complexities of getting AI right?

LL: One important factor to consider is the knowledge of the AI’s applications in the respective organization and the steps needed to enforce this change from ground up. Through conversations with our clients, we have learned that implementing new solutions is often a gradual process. Diving into AI without a plan in place can result in a painful journey and disappointment. We have had clients that explored pilot projects with other service providers that had yielded zero results. 

Many industrial firms we spoke to used systems that were old and extremely restricted, and such systems may have posed difficulties for AI integration. More importantly, they had had previous difficulties identifying the ROI in AI adoption, which is why they had become even more reluctant to move quickly.

In my opinion, the strategy should be to start with small pilot projects with potential vendors to evaluate the effectiveness and actual ROI of a proposed solution. Once they are convinced, they can move on to scale the AI application across the organization.

It also helps to speed up the process if the new solution can be easily integrated into existing legacy systems.

DigiconAsia: What can APAC learn from other regions’ AI adoption and maturation experience?

LL: AI adoption is continuing to rise globally, with companies in emerging markets leading the way. According to one study, Indian firms surveyed had the greatest adoption rate, followed by those in APAC. These firms have shared that engaging in both core and advanced AI practices—such as approaches to building and deploying AI—had helped them get greater returns and achieve higher efficiency. 

Despite close to half of manufacturing and supply chain industries lagging in AI adoption, I believe APAC’s AI adoption rate is relatively good in comparison to other regions: although there is still room for improvement.

Based on my experience, the region is more conservative towards adopting new technology. Coupled with the aforementioned key considerations, including the lack of infrastructure and talent, this results in a much slower adoption rate.

It can help if organizations here can adopt a “fail fast and learn fast” approach to new technology so they can move things along rather than dwell on its setbacks. They should keep in mind that there can always be improvements and iterations along the way.

DigiconAsia: Can every organization benefit from adopting AI at any stage of its digital readiness (or non-readiness)? Does adopting AI mean that an organization must also reach maturity within certain deadlines?

LL: We often tell our clients: “the best time to start was yesterday. The next best time is today.” Any organization can implement AI at any point of its digital readiness, but fundamental IT infrastructure and internet access is a must. 

AI adoption and innovation is a journey, not an end-point. From my point of view, there is no deadline as there will always be new developments that come along the way. What organizations can do is to get started as soon as possible and continuously learn and evolve in tandem with global developments in technology. 

DigiconAsia: From your international experience with MNCs and SMEs, can you share some unique insights with readers in APAC on AI adoption and any other related challenges and indigenous socio-economic issues that we should overcome or circumvent?

LL: One of our clients is in the maritime space, and they are using AI to detect equipment failure and automate manual inspection workflows to improve overall productivity.

The biggest challenge we encountered was in reducing false positives from their previous sensor system by 300%, allowing them to achieve greater efficiency and cost savings. This helped to boost the ground staff’s confidence in technology, and they were able to see the benefits for themselves, based on their faster speed of work. 

I understand that, while AI has been hailed as an approach to enhancing human society, it also has an impact on socio-economic relationships. Concerns about AI’s ability to imitate and replace human labor at a reduced cost, and its impact on worker welfare have been well expressed. Our previous interactions with ground workers have highlighted that they are likewise concerned about being replaced by AI. I empathize deeply, especially when it concerns livelihood. 

But we believe that AI is inevitable, and we take on that responsibility to educate our clients. What we are aiming for is to improve their productivity and reduce manual work so that workers are free to upskill and take on more strategic and value-adding roles or pursue other things in life—and not have to worry of being replaced. 

Think about the industrial revolutions that have happened before this. Steam engines and mass manufacturing pushed mankind forward and created jobs that never existed before. Now, gone are the days of queueing up to withdraw cash from the bank, thanks to the invention of ATMs. These machines pushed bank tellers to upskill and provide better insights and value to the sector. The point here is that each revolution brings about new opportunities.

DigiconAsia thanks Leon for his analytical insights.