Grappling with AI investments

MALAYSIAN companies are often said to be lagging in digital transformation, which could impact their competitiveness.

The discussion gets more animated if artificial intelligence (AI) is the subject. Companies are encouraged to embrace AI as it is expected to revolutionise every aspect of business, and those that fail to adapt may face significant challenges, even closure.

This is why today, companies around the world, Malaysia included, are grappling over how to invest in AI and more importantly, how much to invest. There are signs that the return on investment (ROI) from AI investments (more specifically, generative AI) are not coming to fruition.

A number of top US companies are stuck in their pilot phases, hindered by the lack of high quality data and expertise.

In Malaysia, some companies and government agencies have launched their pilot AI programmes with some degree of success.

In Malaysia, AI is on the government’s agenda as the country plans to become an AI hub. We have the National Artificial Intelligence Office (NAIO), which will coordinate AI development and governance nationally.

There are tax incentives such as deductions for AI-related training, development, and research and development, as well as simplified hiring processes for skilled foreign AI professionals. Budget 2025 has allocated RM50mil specifically for AI education and RM10 mil for establishing NAIO.

While these efforts are a boost, the dilemma remains for some company board members to decide if their company should make a hefty investment in AI. As one board member describes it, “We cannot tell if it’s wholly justified or if this will be another Y2K moment”.

Recall that the Y2K issue, or Year 2000 problem, was a computer bug rooted in how dates were formatted in early computer systems. Companies and governments worldwide spent hundreds of billions of dollars to fix and prepare for Y2K. Alas, many are believed to have overspent.

Consulting firms and contractors made tonnes of money from selling risk-mitigation services. They were essentially the shovel sellers in a gold rush.

It is interesting to note that despite Malaysia’s push for AI, global tech giants are setting up shop here. The bright side is that these companies are promising to train locals to come up to speed with AI, with the ultimate dream of building AI applications in Malaysia.

The large presence of these big tech companies also makes their offerings, installations and maintenance easily accessible to Malaysian corporations.

But are they driven here more to sell their wares, considering how much emphasis is being placed on AI here by the government? After all, aren’t these companies actually the shovel sellers in the AI journey?

Incidentally, this is related to the data centre (DC) boom here. Malaysia has become a fertile ground for tech companies that want to grow their DCs.

Affordable and widely available land, electricity, water and connectivity are the draws, but debate has persisted about the pros and cons.

Malaysia’s Guidelines for Sustainable Development of Data Centre coupled with upcoming specific guidelines on power and water efficiency related to DCs should help mitigate risk factors associated with the DC boom.

One way for corporations to handle AI investment would be to tackle it one project at a time, which each project having very clear use case rationale, and secondly, using third party providers to do the heavy lifting but with fees for that only on a rental basis, similar to how software these days is mostly sold.

This way, companies needn’t hire full-time AI experts, which is costing almost too much, notes one chief executive.

Companies ought to be savvy enough to first find their relevant use cases that solve a problem (or a problem which could potentially crop up as competitors use more AI) and then find the solution provider to solve it on a pilot basis.

With positive outcomes, that project could either be replicated in-house or entirely carried out by third parties for a monthly fee. This will limit corporations’ capital expenditure in their efforts to find that elusive “AI ROI”.