The cost of waiting.
Why delay is not a neutral choice.

Malaysian executives aren’t debating whether to invest in AI anymore. They’re watching competitors move, approving pilots, attending briefings, and still not committing to the one thing that determines whether any of it pays off: getting their teams genuinely capable of using it.

Waiting feels like a reasonable position. You’re not saying no. You’re saying not yet. But in a compounding environment, “not yet” and “never” produce the same outcome over time. The difference is just how long it takes to see it.

What the numbers actually show

BCG’s 2025 global study of more than 1,250 companies is the clearest picture available of where businesses actually stand. Three data points capture the scale of the gap.

Most companies in the 60% don’t think of themselves as laggards. They think of themselves as careful. They have AI tools subscribed to. They’ve had the strategy conversation. They have a pilot or two running. But activity isn’t the same as capability, and the companies in the 5% aren’t just doing more. They’re building differently, and they’ve been building longer.

The compounding mechanic

This is where the cost of waiting gets specific.

AI leaders are already generating returns from their investments, and they’re reinvesting those returns immediately: more structured training, stronger workflows, deeper capability across more of their workforce. Each cycle builds on the last. The result is that with every quarter that passes, the gap between leaders and companies still deciding doesn’t hold steady. It grows.

BCG describes this directly as a “vicious cycle” for laggards. It’s worth sitting with what that means. A company that’s behind today doesn’t stay the same distance behind while it gets ready to move. Leaders are accelerating. Laggards are standing still. By the time a laggard actually begins their program, they’re chasing a target that’s already moving faster than when they started.

The financial numbers make this concrete. Companies BCG identifies as future-built already generate 1.7 times more revenue growth and 1.6 times higher EBIT margins than those producing little to no AI value. They plan to spend 26% more on technology in the year ahead, and allocate 64% more of their IT budget specifically to AI. The gap widens not despite that investment, but because of it.

BCG’s own warning is direct: catching up gets harder with each passing week, not just each passing year. The compounding doesn’t pause while you finalise your roadmap.

Waiting feels like caution. In a compounding environment, it’s the most expensive decision on the table.

Where this plays out at the team level

Pull the lens from company-level financials to the people inside those companies, and the same logic holds.

A team member who’s been working alongside AI tools for 12 months hasn’t just acquired a skill. They’ve built judgment: where AI is genuinely useful, where it isn’t, how to structure prompts for their specific work, which outputs need careful human review, and which workflows have fundamentally changed. That isn’t a certification you can issue to close the gap. It’s accumulated experience, and it doesn’t transfer quickly.

The distance between a team that started 12 months ago and a team starting today isn’t a 12-month skills gap. It’s a judgment gap. And judgment compounds the same way returns do.

Future-built companies upskill more than 50% of their workforce on AI. Companies generating little to no value manage 20%. That gap isn’t just about headcount trained. Companies at 50% are building institutional capability: shared knowledge, adapted workflows, teams that collectively improve as they work. Companies at 20% have pockets of individual skill with no infrastructure to scale it across the organisation.

The productivity difference is measurable. A 2024 Forbes analysis found that employees using AI tools report up to a 40% increase in productivity in relevant workflow areas. At team scale, sustained over 12 months, that gap in output becomes structural. It isn’t fixed by hiring one AI-literate person into a team that doesn’t share the capability.

The window is still open

This isn’t an argument that it’s already too late. It isn’t.

Asia-Pacific companies already allocate the highest share of IT budget to AI of any global region, at 5.2%, and APAC expects a 10% revenue increase from AI by 2028, ahead of both North America and Europe. Malaysia sits inside that momentum, and unlike most of the region, it’s backed by real money sitting in actual government accounts.

Budget 2026 backs it up too: nearly RM20 million for the National AI Office, RM53 million for the Malaysia Digital Acceleration Grant, and a 50% tax deduction for MSMEs on AI training certified by MyMahir.

But the real lever is HRD Corp. It’s named alongside TVET and GiatMARA as one of the government’s vehicles for expanding AI skills this year, and every company with 10 or more employees already pays into it, 1% of monthly wages, automatically. AI literacy training is claimable against that levy right now, under the SBL-Khas scheme.

HRD Corp even imposes a 15% deduction on unused balances over RM50,000, a policy that wouldn’t exist if most companies were actually spending what they’re allocated. That’s the real gap underneath the gap: the funding and the policy are already pointed at exactly this problem, and most organisations still haven’t picked it up.

The more interesting question isn’t whether your organisation will eventually invest in workforce AI capability. Nearly every company will. The question is what the gap looks like by the time you do, and whether the distance between you and your fastest competitors is still a sprint or has quietly become something structural.

The problem isn’t tools

Companies that close this gap won’t do it by buying better software. Every competitor has access to the same tools. The consistent difference between the 5% and the 60% isn’t the technology they use. It’s whether the people using that technology actually know how to use it well.

That’s a training problem. And unlike a technology problem, it doesn’t have a procurement solution. You can’t buy your team 12 months of applied experience. You can only start building it.

The cost of waiting isn’t visible on any balance sheet today. It shows up in revenue growth you didn’t capture, in productivity your competitors extracted while your team was still doing things the slow way, in the compounding distance between where you are and where the leaders already are. It’s quiet. It’s structural. And every week, it gets a little harder to close.

That’s the real gap, and it isn’t about which tools your team has access to. It’s whether they can actually use them well, and keep up as the tools keep changing. SkillTrainer AI exists to close exactly that distance.

We’re building both sides of it: the AI solutions your team will use, and the literacy training that gets them genuinely capable of using them. We’re HRD Corp registered, and our courses are claimable. The compounding starts on day one. The question is which day that is.