YOU may have seen the viral video of a professor screaming at his students at the top of his lungs. “Do you think I can’t tell that you’re all using ChatGPT?” he yells, throwing their assignment papers into the air in rage. “How immature,” we judge silently.
And yet, when we take a step back, don’t we have the similar subconscious fear? Who amongst us has wondered if employees are discreetly using artificial intelligence (AI) tools to think, strategise and draw conclusions?
And if our teams are doing that on their own accord (albeit in silos), might the system be intelligent enough to piece it all together – and have a better and truer view of the company than we do?
If so, we may well be dragged (against our will) into the world of shadow AI: when raw and confidential data is fed onto the public domain, potentially falling into unknown hands, and bringing security risks we are vaguely aware of.
Are we overreacting? Surely not.
Invisibility = Inefficiencies
For all its touted benefits, it is estimated that AI adoption only reached 27% of Malaysian businesses (or 2.4 million in total) in 2024.
This low level is surprising in this high-performance era. A study by Microsoft shows that 61% of leaders demand productivity increase. The survey indicates that 83% of the Malaysian workforce feels overworked, and lacks time and energy to complete tasks.
The proposition of AI is precisely to resolve both dilemmas: create a seamless pro- efficiency environment, and provide the time windfall for resource-stretched employees to refocus on things that matter. Yet, data shows that two of three companies in Malaysia have chosen to wait. The operational drag is already at work.
The cost of waiting is tremendous. Manual inefficiencies have a compounding effect. Employee burnout may turn into attrition, escalating recruitment costs and disrupting team dynamics.
Customer expectations might move faster than internal teams’ awareness and rate of change. And competitors may gain traction.
De-risking strategy
Digital transformation is gradual. Our client records show that companies very rarely embark on a once-for-all encompassing systemic change, but rather, initiate a manageable level of structured high-impact automation exercises. This de-risks business disruption.
In fact, these exercises often lead to repeat undertakings in the organisation, as it scales and/or seeks to mitigate new realities and challenges, such as cybersecurity.
Moreover, returns on investment (ROI) are palpable and measurable – expressed in clearer visibility and accountability, time-savings, higher output, optimised resource planning and the like. Quite the opposite of shadow AI, real AI brings activating insights to light.
A case in point is confectionery and bakery chain Lavender, whose managers witnessed time windfall first-hand after seamless dashboards eliminated manual efficiencies across its 24 outlets. Not only did managers have visibility of operational trackers, it also reduced their production planning time by 50%, from two hours per week to just one.
Crucially, managers reinvested that extra time into studying trends per outlet, monitoring performance of new product launches, focusing on customer engagement, and improving retail experience; the very elements that keep companies ahead.
Take also personal care manufacturer Bodibasixs whose monthly revenue jumped 25% after it consolidated siloed data and accelerated analytics, including saving time on human resource reporting. Plans are underway to enact similar outcomes in the operations and sales divisions.
Malaysia’s AI strides
The Malaysian government has reinforced its commitment to making AI a national priority, as Budget 2026 outlines further infrastructure development such as the RM2bil investment to build a Sovereign AI cloud infrastructure.
Interestingly, the Sovereign AI cloud will increasingly require data to be processed governed within this sovereign framework to ensure security and compliance with the Digital Trust and Data Security Strategy. Companies’ failure to comply creates a barrier to high-impact projects and incentives, hindering the future AI economy.
The government has also set up the National AI Office (NAIO) to coordinate AI policies, standards, and support across public and private sectors.
Correspondingly, data players have shown commitment in empowering Malaysia’s AI agenda. In May last year, Microsoft launched its Malaysia West cloud region, an AI-ready hyperscale cloud offering organisations access to scalable and resilient cloud services with in-country data residency, high levels of security and lower latency.
Google has pledged US$2bil to build a data centre and Google Cloud region in Selangor. Amazon Web Services is assisting SMEs with digitalisation through machine learning and big data analytics.
Infrastructure aside, Malaysia is also cognisant of the need to have safeguards. Reflecting this, Malaysia championed the Asean AI Safety Network last October to develop a common framework for AI governance as a safeguard against risks, even as AI innovation marches on.
All-in, all-win
It is starkly clear that the AI-first strategy is far beyond a mere IT upgrade. It strengthens the foundation of the company and enables everyone to contribute meaningfully to the national agenda.
In the new year, let AI not be an acronym for “active inaction”, but to “activate insights” to advance your growth incline.
