In Singapore, AI is everywhere – except in the work that matters

Across corporate Singapore, artificial intelligence has arrived in all the expected ways.
Policies are being drafted. Enterprise licences are signed. Employees are sent for “AI 101” workshops. Adoption metrics – logins, usage rates, tool access – often look reassuringly strong.

Yet beneath those numbers, a different reality is emerging: most employees still don’t know how to use AI in ways that genuinely change how work gets done.

A recent report based on a survey of 5,000 knowledge workers at large companies in the US, UK and Canada offers a stark picture that will feel familiar in Singapore’s regional headquarters and local corporates alike. AI isn’t failing because people refuse to use it. It’s failing because they don’t have the right kinds of use cases, support or expectations.

For HR leaders here, this is no longer a technology issue. It is a skills, work-design and management issue – and it sits squarely in HR’s domain.

The AI proficiency bar has moved – and most workers are stuck behind it

Until recently, “AI proficiency” meant something fairly basic:
Do your people know what AI is, how to use it safely, and how to write a decent prompt?

Over the past year, companies have invested heavily in these foundations. Employees have learnt to use AI to summarise emails, rewrite messages and answer simple questions. They understand the basics of privacy and bias. On paper, this looks like progress.

But going into 2026, the bar has moved.

Proficiency now means something more operational: can employees incorporate AI into meaningful, value-adding tasks every week? Not as a one-off shortcut, but as a consistent part of how real work is done – in sales funnels, hiring processes, service operations, compliance and analysis.

By that standard, the report’s findings are sobering:

• 97% of the workforce are using AI poorly or not at all.
• Only 2.7% qualify as “AI practitioners” – people who have embedded AI into their workflows and see significant productivity gains.
• Just 0.08% are considered true “AI experts”.

Most employees sit in two large buckets:

• 69% are “AI experimenters”: they use AI for very basic tasks such as summarising meeting notes, rewriting emails and getting quick answers.
• 28% are “AI novices”: they either don’t use AI, or tried it a few times and stopped.

The effect shows up clearly in time saved:

• 24% of workers report saving no time at all with AI.
• Another 44% save less than four hours per week.
• Only a small minority reach higher impact: 23% save two to four hours; 18% save four to eight hours; 8% save eight to twelve; and just 6% say they save 12+ hours a week.

For Singapore employers quietly planning for AI to free up a day or more per week per employee – to offset rising manpower costs and tight talent markets – these numbers should prompt some hard questions.

The “use case desert”: people can prompt, but don’t know what for

Executives often assume the solution is more training. The data points to a different bottleneck:
employees don’t know what to use AI for in the context of their actual jobs.

The report describes a “use case desert”:

• 26% of workers say they do not have a work-related AI use case at all.
• 60% say their existing use cases are beginner-level.
• After analysing 4,500 reported work-related use cases, the authors judged only 15% as likely to generate measurable ROI for the business.

In total:

• 85% of knowledge workers have beginner or no AI use cases.
• 25% say they never use AI for work.
• 40% say they would be fine never using AI again.

Employees understand that AI can summarise a report or polish an email. What they do not see clearly is how AI can change the way a sales team manages pipelines, how a shared services centre handles repetitive requests, or how an HR function screens, assesses and develops talent.

This is not a skills issue in the narrow sense. It is a work-design and role-clarity issue – the kind HR is built to solve.

What employees are actually doing with AI

Look at the “most valuable” use cases workers report, and the gap between promise and reality becomes obvious.

Top work-related AI use cases, by share of knowledge workers:

1. Google search replacement – 14.1%
2. Draft generation – 9.6%
3. Grammar and tone editing – 5.7%
4. Basic data analysis – 3.8%
5. Code generation – 3.3%
6. Ideation and brainstorming – 3.2%
7. Meeting support (e.g. notes) – 2.7%
8. Document summarisation – 2.0%
9. Learning and skill development – 1.6%
10. Task and process automation – 1.6%

When grouped by category, the most popular are:

• Research – 19.6%
• Writing – 18.1%

But both are mostly used in basic ways: one-off copy suggestions and simple informational searches.

More transformative categories – data analysis (6.6%), code (6.2%), task efficiency/automation (4.9%), customer service (2%), strategy (7.1%) – are comparatively underused.

The report’s own assessment is blunt:

• 59% of reported use cases are basic task assistance.
• Over 25% had no relevant role in larger processes or workflows.
• Only 2% of use cases were judged to be advanced.

In effect, AI is functioning as a convenience layer – a smarter spellcheck and search engine – rather than a serious productivity engine.

For Singapore employers under constant pressure to do more with lean teams, this is lost opportunity at scale.

A leadership optimism gap: policies on paper, frustration on the ground

The report also exposes a sizeable disconnect between how senior leaders see their AI deployments and how the rest of the organisation experiences them.

Among C‑suite respondents:

• 75% say they are excited about AI’s implications.
• 94% say they trust AI’s contributions.
• 57% use AI for work daily; only 2% don’t use it for work at all.

They also overwhelmingly believe their companies have the right infrastructure in place:

• 81% say there is a clear, actionable AI policy that effectively guides use.
• 80% say tools exist with a clear access process.
• 71% say there is a formal AI strategy.
• 66% feel encouraged to experiment and build their own AI solutions.
• 48% believe there is widespread adoption with open sharing of use cases and best practices.

Ask individual contributors – employees without direct reports, who handle much of the day-to-day work – and the picture looks very different:

• Only 28% agree there is a clear, actionable AI policy – a 53‑point gap from the C‑suite.
• 39% say tools exist with a clear access process (versus 80% of executives).
• 32% say there is a formal AI strategy (against 71%).
• 25% feel encouraged to experiment and build solutions (versus 66%).
• Just 8% believe there is widespread adoption with open sharing of use cases (versus 48%).

For HR leaders in Singapore who must brief optimistic executive teams while hearing frustrations from the ground, these numbers will feel familiar. They show not only a communication gap, but a risk of complacency at the top and cynicism at the bottom.

Individual contributors: doing the repetitive work, getting the least support

The inequity is not just perceptual; it’s structural.

Individual contributors (ICs) are the least likely to benefit from their company’s AI resources – despite being the ones performing much of the repetitive, automatable work.

On access to tools:

• Only 32% of ICs say they have clear access to AI tools.
• That rises to 57% for managers, 72% for directors and VPs, and 80% for C‑suite leaders.

On training:

• Just 27% of ICs have received company AI training.
• For managers, it is 48%; for directors 68%; for VPs 70%; and for the C‑suite 81%.

On reimbursement:

• Only 7% of ICs are reimbursed for AI tools.
• Managers: 26%; directors: 44%; VPs: 33%; C‑suite: 63%.

Unsurprisingly, ICs are more likely to feel anxious or overwhelmed by AI, less likely to trust it, and less likely to say it is having a transformative impact on their work.
Manager support is weak: only 7% of ICs say their managers expect daily AI use, and only about a third receive consistent encouragement to use it. Since May 2025, manager support for IC AI use has actually declined by 11%.

For Singapore companies with regional hubs, shared services centres and large frontline workforces, this is a critical misalignment. The people whose time is most “freeable” by AI are the very ones least supported to use it.

Where industries and functions stand – and what that means in Singapore

Although the survey covers North America and the UK, its patterns map readily onto Singapore’s sector mix.

By industry, AI proficiency scores (out of 100) are:

• Technology – 42
• Finance – 36
• Consulting – 35
• Manufacturing – 34
• Media – 33
• Real estate – 32
• Food & beverage – 29
• Education – 29
• Healthcare – 28
• Retail – 27

The leading sectors are more likely to have a formal AI strategy, clear policies and good tool access. Lagging sectors – including healthcare, education and retail, all significant in Singapore’s economy – are more likely to lack a strategy altogether or have unclear policies.

By function, proficiency looks like this:

• Engineering/Tech – 41
• Strategy – 39
• Business development/Sales – 37
• Human resources – 37
• Marketing – 36
• Finance/Legal – 35
• Product – 34
• Operations – 32
• Customer service/Support – 27

Even within higher-scoring functions, obvious opportunities are being missed:

• 54% of engineers don’t use AI for writing or debugging code, scripts or formulas – one of the clearest use cases for their role.
• 56% of marketers don’t use AI for creating first drafts of content.
• 87% of product managers don’t use AI for creating prototypes.

For Singapore organisations positioning themselves as regional digital and innovation hubs, these gaps speak directly to competitiveness. The tools are there; the high-value use cases often are not.

Training and investment: moving the needle, but not enough

The report does recognise progress. Since March 2025:

• Access to a formal AI policy is up 17%.
• Clear guidelines for AI usage are up 16%.
• Investment in AI tools and platforms is up 2%.

These investments do have impact:

• Employees at companies with a company AI strategy are 1.6× more proficient than those without one.
• Employees with access to AI tools are 1.5× more proficient than those with no access.
• Employees who have been trained on AI are 1.5× more proficient than those who have not.
• Employees whose managers expect AI usage are 2.6× more proficient than those whose managers discourage it.

And yet, even in these “higher proficiency” groups, most workers are still just “experimenters”. The average worker who has undergone AI training scores only 40 out of 100 in AI proficiency.

The reason is straightforward: most programmes are still focused on access, safety and prompting. They teach people how to use an LLM and how not to leak data. They do not systematically teach them how to identify workflow bottlenecks, design high-value use cases, and re-engineer processes with AI in mind.

Training is getting employees to the starting line, not across the finishing line.

What Singapore HR leaders should do now

The report’s recommendations for leaders translate cleanly into an HR agenda for Singapore.

1. Stop measuring success by access and adoption alone

If 55% of your workforce uses AI weekly but only 15% have value-driving use cases, your adoption metrics are misleading.

HR can push for metrics that track:

• Time saved per role and function.
• The quality and maturity of use cases (basic vs intermediate vs advanced).
• The share of core workflows that have been redesigned with AI.
• Concrete business outcomes: cycle times, error rates, customer satisfaction, revenue or cost impact.

2. Treat use case development as a core competency – not an individual side project

Employees are not stuck because they can’t prompt; they’re stuck because they don’t know what problems AI can solve in their role.

HR can:

• Build function-specific use case libraries (for HR, finance, operations, customer service, sales, etc.).
• Create role-based playbooks showing what “basic, intermediate, advanced” AI use looks like in each job family.
• Make use case development an explicit part of managers’ responsibilities and performance conversations.

3. Bridge the individual contributor gap quickly

ICs are doing much of the repeatable work; they should not be last in line for tools and support.

In Singapore’s context of manpower constraints and high labour costs, HR can:

• Standardise AI tool access for eligible roles, not just seniority levels.
• Create clear policies for reimbursing or centrally funding approved AI tools.
• Mandate that every manager identify and track at least three AI use cases per direct report.

4. Redesign training around workflows, not just tools

Safety and prompt-writing are necessary but insufficient.

New programmes should help employees:

• Map their own workflows and identify repetitive or slow steps.
• Experiment with AI to eliminate, automate or accelerate those steps.
• Evaluate AI outputs critically and build habitual, weekly usage into routines.
• This is where HR’s capability in learning design and change management becomes central.

​​​​​​​5. Close the executive awareness gap

When C‑suites see AI as a success while ICs see minimal impact, data visibility – and trust – suffer.

HR can:

• Run regular skip‑level sessions focused specifically on AI barriers, not just high-level strategy.
• Ask executives to shadow employees as they attempt to use AI in real work (not just in demos).
• Provide dashboards that surface where and why AI use is shallow, stalled or failing – not just where it is succeeding.

​​​​​​​6. Accept that the AI proficiency bar will keep rising

The gap between experimenters and practitioners will widen as AI capabilities advance.

HR should:

• Build continuous learning infrastructure, not one-off workshops: internal academies, communities of practice, mentoring and certification.
• Embed AI proficiency into competency frameworks and career paths across functions, not only in IT or data.

The real test for Singapore’s HR leaders

The report’s deeper message is simple: AI transformation in the workplace is not primarily about technology deployment. It is about redesigning how work is done – task by task, workflow by workflow, role by role.

For HR leaders in Singapore, that places AI squarely within your mandate: skills, culture, work design, managerial expectations and measurement.

Globally, adoption numbers can look impressive. ChatGPT reports nearly 900 million monthly users; 56% of Americans say they use AI. Yet underneath, 85% of the workforce still lacks a value-driving AI use case, and 25% do not use AI for work at all.

For Singapore organisations, the key question in 2026 is no longer, “Do our employees have access to AI?” or “Have they completed the training?”

It is: Have we helped our people turn AI from a clever assistant at the edges of their day into a core part of how value is created – and are we honest about the answer?

That is not a question IT can solve alone.

It is a question for HR. And it is becoming urgent.