The real AI race is for deployable talent, not models

SINGAPORE is entering a new phase of its artificial intelligence journey, where the pressing question is how to deliver its benefits at scale.

Our digital economy is worth S$128.1 billion, accounting for 18.6 per cent of gross domestic product. Growth is increasingly driven by digitalisation across sectors beyond just technology.

Nearly three in four Singapore workers reported using AI tools at work, with small and medium-sized enterprise (SME) adoption rates tripling from 4.2 per cent to 14.5 per cent and non-SME adoption rates rising from 44 per cent to 62.5 per cent, based on the latest Singapore Digital Economy Report.

Yet as AI moves from experimentation into deployment, a more fundamental constraint is emerging.

Many firms struggle to drive business transformation despite strong adoption and investment. A 2025 McKinsey survey found that around 60 per cent of South-east Asian companies reported less than a 5 per cent earnings impact from AI, with nearly one in five seeing no meaningful effect at all.

The challenge is no longer just access to models or compute alone. It is execution – the ability to translate AI into real work, consistently and at scale.

Moving beyond AI literacy to AI bilingualism

As AI adoption accelerates, literacy remains necessary – but it is no longer sufficient.

Many organisations can deploy AI, but far fewer can operationalise it across workflows, decisions and day-to-day business processes. This is where most AI efforts stall – not at the point of innovation, but execution.

What is increasingly needed is AI bilingualism: the ability to understand AI’s capabilities, limitations and risks, and apply it effectively within one’s domain.

With AI, as with any language, fluency requires more than vocabulary. It requires mastery to solve problems, improve decisions and deliver outcomes in context. Beyond specialists, it is critical to ensure the technology can also be applied by non-tech professionals driving innovation across sectors.

AI bilingualism is therefore more than a workforce skill; it is the missing execution layer.

Many organisations struggle to move beyond pilot projects because they lack people who can translate AI into business outcomes.

For workers, the same gap creates uncertainty about how their roles will evolve, where they can add value, and if they can keep up with ever-shifting technologies and business priorities.

In short, we must build capability where it matters.

Redesigning roles and workflows

To meet new realities, enterprises must first redesign work and learning.

AI cannot simply be layered onto existing processes. Businesses need to redesign roles to evolve alongside technology, rethink workflows, and support employees through structured, industry-relevant upskilling pathways.

A common but flawed approach with AI is to drive end-to-end automation of tasks with as minimal headcount as possible. In logistics, for instance, fully automated route planning may cut costs in stable conditions, but often falls short when real-world disruptions arise.

A more sustainable approach is to keep humans in the loop, redesign roles, and shift work from manual planning to overseeing outputs, managing exceptional situations and improving system performance over time.

This will be especially important in professions highly exposed to AI, such as accountancy and legal services, where adoption must be matched by workers’ ability to apply AI effectively in their roles.

Building applied capabilities

We are seeing the emergence of a “two-speed workforce”. One half comprises AI-augmented, high-productivity employees. The other is a growing cohort of displaced workers and fresh graduates with limited experience – who feel the pressure to stack market-relevant credentials.

Helping the second group get up to speed requires moving beyond short courses and paper certifications, and towards learning through practice and continuous adaptation.

Workers must learn to interpret AI outputs, be comfortable with unknowns and probabilistic thinking when faced with incomplete data, understand how to apply AI in context, and exercise judgment when systems fall short.

Enterprises can support this by creating environments where employees are encouraged to experiment with AI and apply it in daily work.

Governments and industry partners can expand access to learning opportunities through platforms that bring together policymakers, businesses and practitioners.

Cross-sector convenings such as the upcoming Asia Tech x Singapore can also help workers deepen their practical AI knowledge.

Collaborating across sectors

The public and private sectors must scale AI bilingualism together.

In Singapore, efforts are already under way to support this shift.

The National AI Impact Programme, for example, aims to drive AI adoption across enterprises while building a base of AI bilingual workers.

With an initial focus on accountancy and legal professionals, this programme provides participants with hands-on learning experiences on redesigning workflows with AI, while strengthening their understanding of responsible AI and profession-specific risks.

Such initiatives empower professionals to strengthen AI bilingualism, raising productivity and enabling them to focus on higher-value tasks that require specialised skill sets such as risk analysis and client advisory.

For SMEs, the challenge is knowing how to use AI well. Beyond cost, complexity and training accessibility, many need clearer guidance to identify where exactly AI can make a real difference.

For more digitally mature enterprises, the challenge is integrating AI into core workflows and decision-making. Leaders play a key role in setting direction, aligning teams and driving adoption across the organisation.

As AI reshapes the technology sector, enterprises must also support their tech workforce in moving into higher-value roles. Engineers and technical professionals will increasingly need to build and orchestrate more complex AI-driven workflows – and engage the wider workforce to sustain innovation.

AI bilingualism will define competitiveness

Singapore has successfully navigated major economic transitions before, each requiring adaptation, investment in skills and cross-sector collaboration.

This time, competitiveness will depend not just on adopting technology but on how willing we are to learn a new “language”, adapt how we work, and apply AI effectively across roles and industries, at scale.

The writer is cluster director of the human capital cluster at Infocomm Media Development Authority of Singapore