Which Companies Offer AI Upskilling Programs in Singapore? (2026 Complete Enterprise Guide)

A guide for chief learning officers, chief human resources officers, transformation leaders and chief information officers evaluating AI upskilling providers in Singapore: who is in the market, how they compare, what the government will fund before the November deadline, and why most programmes still fail to change how work gets done.

The short answer

The right choice depends less on the provider's brand than on what you are actually buying. A course teaches a tool. A programme changes a workflow. A capability build changes how the organisation decides, governs and measures. Most enterprises buy the first and expect the third.

AI upskilling providers in Singapore compared

The table below covers the providers an enterprise buyer in Singapore is most likely to shortlist. Attributes are drawn from each organisation's own published information as at July 2026; formats, fees and course lists change often, so confirm details directly with the provider before committing.

Two omissions are deliberate. Prices are not listed because published fees change too quickly to be reliable in a guide, and the funding position matters more than the sticker price; the funding section below covers that. Ratings are not listed because no independent rating system covers this market consistently.

How to read the market: the Six-Category Provider Map

1. National and government-backed programmes

AI Singapore's apprenticeship model deep-skills local AI engineers on real industry projects. IMDA's TeSA has, on its own reported figures, placed more than 24,300 locals into technology roles and upskilled more than 440,000 individuals with technology skills to date. Under the National AI Impact Programme, a Digital Leaders Accelerator Bootcamp is being introduced for business leaders. Best for building deep technical talent at low cost. The trade-off is that national programmes serve national goals, not your specific operating model.

2. Universities and institutes of higher learning

NUS-ISS, SMU Academy, NTU and SUTD Academy carry university credentials and academic rigour, with stackable structures built for working professionals. Best for recognised credentials and technical depth for specialists. The trade-off is pace and specificity: a certificate changes what an individual knows, not what a team's Monday morning workflow looks like.

3. Global consultancies and systems integrators

Accenture, Deloitte, PwC, EY, KPMG and IBM typically deliver capability building as one workstream inside a larger transformation, scoped and priced accordingly. This model is built for multi-country programmes where training, technology and process change move under one contract. The trade-offs are structural to the model rather than criticisms of any firm: capability building carries consulting cost structures, and continuity after the engagement closes depends on what was explicitly contracted.

4. Technology vendors and platforms

Microsoft, Google, Amazon Web Services and Nvidia provide deep, often free, learning aligned to their own platforms, alongside horizontal platforms such as Coursera and LinkedIn Learning. Best for technical certification on a stack you have already committed to, at scale. The trade-off is inherent: vendor curricula teach the vendor's tools, and governance framed for Singapore's regulatory context is not their job.

5. Continuing education and training providers

NTUC LearningHub, General Assembly, Vertical Institute, Heicoders Academy and Lithan compete on accessible, funded, short-format courses, and between them cover most of the subsidised course volume in this market. Best for foundational literacy across a large population quickly and cheaply. The trade-off is that short courses build individual skills; workflow redesign and adoption are left to you.

6. Specialist AI transformation firms with an academy

A smaller group combines consulting with structured certification, so diagnosis, programme design and adoption support come from one accountable team. MASSIVUE sits here, alongside boutique advisories in the same space. Best for organisations that want the workflow redesign and the training from the same people. The trade-off is brand: buyers who need a global name on the invoice for board comfort will look to category three.

Which category fits your situation

The 2026 funding position, and two deadlines

The two dates. The current SkillsFuture Enterprise Credit expires on 30 November 2026, and for training to qualify, the course run must be completed on or before that date; unused credit is forfeited. A redesigned SFEC launches on 1 December 2026 under the Enterprise Workforce Transformation Package, giving eligible employers a fresh S$10,000 through an online wallet usable upfront at enrolment (SkillsFuture and GoBusiness).

The administrative change. SkillsFuture Singapore and Workforce Singapore merged on 1 July 2026 into the Skills and Workforce Development Agency (SWDA), a single statutory board jointly overseen by the Ministry of Manpower and the Ministry of Education (MOM). Existing schemes continue, but portal names and claim routes are consolidating through the year, so check the current process at the point of enrolment.

What is AI upskilling?

This guide uses a five-step AI Capability Ladder to describe progress: awareness, literacy, tool fluency, upskilling (workflow redesign), and capability (self-sustaining). Most enterprise programmes stop at step three. The value is on steps four and five.

Why AI upskilling matters in Singapore in 2026

Adoption outran capability. The Infocomm Media Development Authority (IMDA) reported that AI adoption among small and medium-sized enterprises tripled from 4.2% in 2023 to 14.5% in 2024, while adoption among larger firms rose from 44% to 62.5% (Singapore Digital Economy Report 2025). The tools arrived. The redesigned roles largely did not.

The state set a people target. The National AI Impact Programme, launched at the Committee of Supply debates on 2 March 2026, will support 10,000 enterprises over three years and back 100,000 workers to become AI Bilingual, meaning fluent in AI and deep in their own professional domain, starting with accountancy and legal professionals (MDDI).

Skills are the stated constraint. In the World Economic Forum's Future of Jobs Report 2025, 63% of employers named skills gaps as the single biggest barrier to business transformation. The same report expects 39% of workers' core skills to change by 2030, with 59 in every 100 workers needing training, of whom 11 are unlikely to receive it.

Against all of that, Boston Consulting Group's Build for the Future x AI 2025 study found only about 5% of organisations had achieved substantial financial gains from AI, and that this group delivered three-year total shareholder returns roughly four times higher than laggards (BCG, February 2026). The gap between adoption and value is a people gap. That is the case for upskilling, and it is stronger than any vendor claim.

Signs your organisation needs AI upskilling

Our AI readiness assessment covers the same ground in a structured form.

How enterprise AI upskilling actually works

The four layers

The manager layer is decisive. In the companies extracting the most value from AI, BCG found 88% of managers actively used AI in decision making and daily operations, against 25% at laggards; the same companies were four times more likely to run structured AI learning programmes with protected time to learn.

Governance

Governance defines what people may do with which data, and who is accountable when an AI system gets it wrong, and in Singapore this is no longer optional framing. IMDA's Model AI Governance Framework for Agentic AI, launched in January 2026, is built on the principle that a human must remain meaningfully accountable, which is a training requirement as much as a policy statement. For financial institutions, the Monetary Authority of Singapore's proposed Guidelines on AI Risk Management explicitly address the capabilities and capacity institutions need for the use of AI, with a 12-month transition period proposed after issuance. Consultation closed on 31 January 2026; the final guidelines had not been issued as at the end of July 2026, so confirm current status with MAS before planning against them. For why governance failures end AI programmes, see Agentic AI Governance: Why Enterprises Are Decommissioning Their Agents.

Change management

Change management determines whether new behaviour holds after the launch campaign ends. We set out the full six-stage lifecycle in End-to-End Change Management in Singapore: A 2026 Playbook, including the sequencing question most organisations cannot answer: how many concurrent changes is each team already absorbing?

Measurement

Measurement decides whether the programme gets funded again. Completion rates measure attendance. The metrics that matter are utilisation, proficiency, workflows in production, and the business outcome attached to each one. BCG calls the speed at which skills are learned, applied and embedded "capability velocity", and treats it as the variable that separates AI leaders from the rest.

Build internally or buy from a provider?

The failure mode on the build side is assigning AI upskilling to a learning management system and a playlist. The failure mode on the buy side is renting a capability and returning it when the invoice ends. The test for any external provider is whether they are contractually building your independence; that is criterion ten on the scorecard below.

What makes a great AI upskilling provider?

Two criteria matter most. Baselining before design (criterion 6), because a measure introduced after the programme cannot show what the programme changed. And the 90 days after the final session (criterion 8), because BCG's consistent finding is that skills stick when learning is embedded into daily work with real tools, real tasks and reinforcement, not when it is taught in isolation.

Why many AI training programmes fail

The failure modes are consistent:

The pattern underneath all seven: organisations buy the 10% and the 20% and skip the 70%. BCG's 10-20-70 finding, drawn from its case work with hundreds of companies, attributes about 10% of AI value to algorithms, 20% to the technology required to implement them, and 70% to rethinking the people component. Generic content is a related trap; we have written separately on why generic AI training fails, department by department, because finance, legal, sales and operations do not share a workflow, a risk profile or a definition of a good outcome. An AI strategy without an operating model fails the same way, for the same reason, as we argue in Why Your AI Strategy Needs an Operating Model.

A worked example: what the numbers should look like

Take a team of 40 analysts whose baseline shows 6 hours per person per week on report assembly, data cleaning and first-draft writing. Suppose a 12-week programme, covering tool fluency plus the redesign of those three workflows, returns 3 of those hours per person per week, verified by utilisation and proficiency data rather than self-report. That is 120 hours per week across the team, roughly 5,500 hours per year allowing for leave. At a fully loaded cost of S$60 per hour, the annual value is about S$330,000, against which you set the programme fee, the absentee payroll cost of training time, and the funding offsets from the schemes above.

Three honesty rules make the model defensible in front of a chief financial officer. Count only hours verified against the baseline, not survey estimates. Deduct the ramp: productivity typically dips before it improves, which BCG's case work shows is precisely when incentives and metrics must be adjusted or people revert. And attribute value to redeployed hours only if you can say what the hours were redeployed to.

How MASSIVUE approaches AI upskilling

Diagnosis first. Engagements start with a baseline: AI readiness, current workflows, risk posture and the skills already present. AI Workforce Transformation covers skills assessment and gap analysis, role design and the certification pathway that follows.

A tiered adoption programme. AI at Work is a three-level pathway (AI 101, 201, 301) with department-specific tracks, built for organisations rolling out AI adoption across 100 to 500 or more employees in a phased sequence rather than a single event. Industry packages exist for manufacturing, banking and financial services, energy, and telecommunications; sector-specific work is catalogued on our industry page.

An operating model, not just a curriculum. Protum is our AI operating model, covering six business capabilities for human and AI integration, from data culture and adaptive structures to responsible intelligence and impact prioritisation. It is what turns trained individuals into a functioning system, and it anchors our enterprise transformation work.

Credentials that stack. MASSIVUE Academy delivers certifications built from micro-credentials, including the Certified Practitioner in AI and Digital Transformation, so learning accumulates rather than expiring with a workshop. MASSIVUE's institutional partners include NTUC LearningHub in Singapore, with whom we develop course materials and IBF Standards Training Scheme accreditation packs for financial sector programmes; Asia e University in Malaysia; and the international certification body TÜV NORD.

Sustainability capability where it is required. SustainAgility is our sustainability framework, and the Academy carries green finance and environmental, social and governance tracks for organisations where AI and sustainability reporting obligations intersect.

On our own numbers. MASSIVUE's published outcomes include a client engagement reporting 89% skills gap closure across 1,200 employees in six months at a technology company, and the firm was named Best AI-Led Business Transformation Consultancy 2026 at the Singapore Business Awards; both are published on our site, and the outcome figures are self-reported from client engagements rather than independently audited. We would apply the same scepticism to any provider's numbers, including ours: ask for the measurement method behind any headline, and read our case studies and results wall with that question in hand.

A note for public sector organisations

The evaluation criteria in this guide still apply, but two adjustments matter. First, data governance requirements are stricter and non-negotiable, so a provider's answer to criterion 7 carries more weight. Second, sustainment usually depends on internal learning functions, so the handover in criterion 10 should be specified in the contract, not assumed. Agencies evaluating providers should verify current procurement routes and any whole-of-government AI training arrangements through their own channels, as these change and are not always public.

A realistic 12-month sequence

The executive checklist

Before you sign with any provider:

• A named executive sponsor with the outcome in their own objectives
• A baseline captured before design begins, not after
• Role-specific content for each affected function, not one deck
• Data and governance rules published before the first session
• Adoption metrics defined, with owners and a review cadence past launch
• A costed plan for the 90 days after the final session
• Funding position confirmed against current scheme rules, and the 30 November 2026 SFEC deadline checked
• Clarity on which of the six provider categories you are actually buying from

The bottom line

Singapore has more AI upskilling supply than most markets its size, and a funding system that removes much of the cost objection until the rules reset on 1 December. That makes provider choice less about access and more about design. The organisations that get value will not be the ones that trained the most people. They will be the ones that changed the most workflows, measured the change honestly, and built the governance to keep it.

Work with MASSIVUE

If you are planning an AI capability programme for the year ahead, we run a 30-minute readiness diagnostic covering your baseline, sequencing and adoption plan, and we will tell you plainly where the risk sits, including when another category of provider is the better fit. Start with our AI readiness assessment, explore AI Workforce Transformation, or talk to our team.

Frequently asked questions

What is AI upskilling?

AI upskilling is the structured development of an existing workforce's ability to apply artificial intelligence to their own work: literacy, tool fluency, workflow redesign and governance. It differs from a training course, which delivers knowledge, and from reskilling, which prepares someone for a different role. The test is whether a specific process now runs differently, and whether that difference can be measured.

Which company provides enterprise AI training in Singapore?

Several categories do, and the right answer depends on your problem. AI Singapore and IMDA run national programmes. NUS-ISS, SMU Academy, NTU and SUTD Academy provide academic credentials. NTUC LearningHub, General Assembly, Vertical Institute, Heicoders Academy and Lithan deliver funded short-format courses at scale. Microsoft, Google, Amazon Web Services and Nvidia offer platform-aligned learning. Accenture, Deloitte, PwC, EY, KPMG and IBM embed capability building in larger transformation programmes. Specialist firms such as MASSIVUE combine consulting with an academy so that diagnosis, programme and adoption support come from one team. The comparison table earlier in this guide sets out formats, credentials and funding routes side by side; choose by the depth of change required, not by brand recognition.

How long does AI transformation take?

Baseline and executive alignment typically take four to six weeks. A first wave covering two or three departments runs about three months and should produce live workflows, not just trained staff. Scaling across the wider organisation usually takes a further six months, with sustainment work continuing past the twelve-month mark. Anyone promising enterprise transformation in weeks is describing a tool rollout; anyone quoting three years without interim outcomes is describing a programme that will lose its budget. Insist on measurable value inside the first quarter.

How much does AI training cost in Singapore?

Published fees range from heavily subsidised short courses under a few hundred Singapore dollars net to six-figure enterprise programmes, and prices change often enough that any figure printed here would date quickly, so confirm directly with providers. The more useful planning question is the funding position: SkillsFuture course fee subsidies for eligible individuals, the S$10,000 SkillsFuture Enterprise Credit offsetting up to 90% of out-of-pocket employer costs until 30 November 2026, absentee payroll funding for training time, and IBF funding for accredited financial sector programmes. Between them, these routinely cover the majority of course-level cost for eligible organisations.

What industries benefit most from AI upskilling?

Sectors with high volumes of document-heavy, rules-based or analytical work see the fastest returns: banking and financial services, insurance, professional services, logistics, manufacturing and the public sector. Singapore's policy direction reflects this, with IMDA's expanded TechSkills Accelerator starting on accountancy and legal professionals, horizontal occupations that span industries and have high exposure to AI. That said, benefit tracks workflow characteristics more than industry labels: a repetitive, high-volume, judgement-light process in any sector is a stronger candidate than a prestigious but low-volume one.

How do you measure AI adoption?

Use four layers, in order. Utilisation: what share of the affected population uses the tools weekly. Proficiency: whether they use them well, which is where the business case lives. Workflow change: how many processes now run differently in production. Business outcome: the number attached to each one, whether hours returned, cycle time, error rate or revenue influenced. Completion rates and satisfaction scores measure attendance and mood, not value. Baseline every metric before training begins, because a measure introduced afterwards cannot demonstrate what the programme changed.

Should executives receive different AI training from employees?

Yes, and the difference is content, not simplification. Executives need the investment logic, the risk and governance decisions only they can make, the workforce implications, and enough personal fluency to be credible. Managers, the layer most often skipped, need to redesign their team's workflows and coach usage; BCG found that in the companies extracting the most value from AI, 88% of managers actively used AI in decision making, against 25% at laggards. Employees adopt what their direct manager visibly uses, not what corporate communications announce.

What AI certifications matter for enterprises?

Credentials that assess applied capability against a recognised framework carry more weight than attendance certificates. In Singapore, courses aligned to national frameworks, and for the financial sector, programmes accredited under the Skills Framework for Financial Services, carry both recognition and funding eligibility. Vendor certifications from Microsoft, Google, Amazon Web Services and Nvidia are valuable for technical specialists on a chosen stack. The more important question is what the certification actually assesses: applied work on real tasks, or recall of course content.

How do you build AI capability across a whole organisation?

Sequence it. Baseline skills, workflows and risk. Enable leaders and managers before the wider population, because adoption follows visible managerial use. Run a first wave in two or three functions using real workflows and real data rules, and measure it properly. Publish governance rules before anyone trains. Scale through a champion network and stackable credentials, then update job descriptions, onboarding and performance conversations so the new way of working becomes the default. Capability is the point at which the organisation improves its own AI work without external help.

What is enterprise AI maturity?

Enterprise AI maturity is the degree to which an organisation can select, deploy, govern, measure and improve AI systems as part of normal operations rather than as projects. Immature organisations run pilots, depend on individual enthusiasts, and cannot state the value delivered. Mature organisations have named accountability for AI decisions, an operating model connecting business and technical roles, adoption data reviewed on a cadence, and a repeatable path from idea to production. BCG's 2025 global study put organisations achieving substantial financial gains from AI at only about 5%, a useful reminder of how rare genuine maturity still is.

How do micro-credentials help with AI upskilling?

Micro-credentials break a qualification into short assessed units that stack toward a full certification. They fit adult working schedules, so completion holds up better than on long programmes; they let an organisation build precisely the competencies a role needs; and they create visible progression, which supports retention in a market where people leave when they stop learning. MASSIVUE Academy uses this structure, with certifications such as the Certified Practitioner in AI and Digital Transformation assembled from stacked micro-credentials.

How do you prepare employees for AI in their roles?

Start before the tools arrive. Tell people honestly what is changing, what it means for their role, and what the organisation commits to in return. Baseline current skills and sentiment so enablement can be targeted. Publish data and usage rules early, because uncertainty about what is permitted suppresses adoption more than lack of skill does. Train on the actual work rather than tool features, give protected time to practise, then update job descriptions and performance conversations so the new expectations are formal rather than implied.

How do you reduce employee resistance to AI?

Treat resistance as information. Most of it comes from three sources: fear about job security, uncertainty about permitted use, and change fatigue from too many concurrent initiatives. Be specific about what AI will and will not replace in each role, and honest where you do not yet know. Publish clear governance so people are not guessing. Count how many other changes each team is absorbing before adding another. BCG's client work adds a fourth source worth naming: professional identity. People who built careers on doing the work can experience directing AI as a loss; effective programmes redefine the role explicitly, showing where human judgement still decides, rather than hoping the discomfort fades.

What role does change management play in AI upskilling?

It is the difference between trained staff and changed work. Training delivers knowledge; change management determines whether the knowledge alters daily behaviour and whether that behaviour survives the programme ending. It covers the case for change, readiness assessment, participatory design, role-specific enablement, adoption tracking and sustainment. BCG attributes about 70% of AI value to the people and process component. Our companion guide, End-to-End Change Management in Singapore, sets out the full six-stage lifecycle with named owners and the evidence each stage was genuinely done.

Can small and medium-sized enterprises benefit from AI upskilling?

Yes, and Singapore's funding position makes the economics unusually favourable. IMDA data shows SME AI adoption tripled from 4.2% in 2023 to 14.5% in 2024, so the peer group is moving. Eligible employers can access the S$10,000 SkillsFuture Enterprise Credit covering up to 90% of out-of-pocket costs until 30 November 2026, alongside course fee subsidies and absentee payroll funding, and the National AI Impact Programme is built to support 10,000 enterprises over three years. The practical advice for smaller firms differs from enterprise advice: pick one workflow that consumes real hours, train the people who own it, measure the result, and only then expand.

What is the difference between AI literacy and AI fluency?

Literacy is understanding: what the technology does, where it fails, what the risks are, what your organisation permits. Fluency is doing: applying the tools competently to your own work, recognising when output is wrong, and knowing when not to use them at all. Literacy can be delivered at scale through short courses; fluency requires practice on real tasks with feedback, which is why programmes that stop at awareness rarely move utilisation.

What does "AI Bilingual" mean in Singapore?

AI Bilingual is the term Singapore's National AI Impact Programme uses for workers who combine AI fluency with genuine depth in their own professional domain, so they can redesign their field's workflows rather than merely operate tools. The programme, announced by the Ministry of Digital Development and Information in March 2026, aims to support 100,000 workers in becoming AI Bilingual, with IMDA's expanded TechSkills Accelerator delivering the first tracks for accountancy and legal professionals from the first half of 2026. For enterprises, it is a useful bar: an AI Bilingual employee is the four-layer model's employee and specialist layers meeting in one person.

How does AI governance connect to upskilling?

Governance defines what people are permitted to do; upskilling determines whether they can do it competently. The two fail together. Singapore has made the link explicit: IMDA's Model AI Governance Framework for Agentic AI rests on humans remaining meaningfully accountable, and a person cannot be accountable for a system they do not understand. For financial institutions, MAS's proposed Guidelines on AI Risk Management address the capabilities and capacity needed for AI use; they were not yet finalised as at July 2026, so verify current status before planning against them.

Should we build AI training internally or hire a provider?

Usually both, sequenced. Buy the diagnosis, the workflow redesign and the first wave from a provider with pattern recognition across organisations; build the internal capability that runs literacy at scale, sustains adoption and extends the programme in year two. The build-or-buy table earlier in this guide gives the five deciding questions. The one non-negotiable, whichever way you lean: any external contract should specify how the engagement ends with you not needing the provider, because dependence is the quiet failure mode of bought capability.

What does "enterprise-safe" AI training mean?

Enterprise-safe training means the programme is designed so that no confidential company data enters unapproved AI systems at any point, including during exercises. In practice that requires four things: training conducted on platforms your IT and security functions have approved; explicit data-boundary rules taught from the first session, not appended as a policy link; exercises built on synthetic or sanctioned data rather than whatever participants paste in; and clarity on where prompts and outputs are logged and retained. If a provider cannot answer where your data goes during their own training, that is your answer.

How many employees should we train first?

Fewer than instinct suggests. Start with the leadership and manager layers, then a first wave of 30 to 80 people in two or three departments chosen because they own workflows with measurable volume. That produces better evidence than 500 people trained thinly, and it gives you real data before the larger budget commitment. The exception is foundational literacy, which can sensibly go wide and cheap early, provided it is understood as a floor rather than the programme. Scale after the first wave shows sustained utilisation, not after it shows good feedback scores.

What should be in an AI upskilling request for proposal?

Eight things. Your baseline data, or a requirement that the provider produce one. The specific workflows in scope, named. Role-based tracks rather than a single audience. Data handling rules, including approved platforms and where training data goes. Assessment and credential detail, including what is actually tested. Adoption metrics with baselines, owners and a review cadence extending past launch. Post-programme support, costed. And a statement of which provider category you are inviting, because comparing a university certificate against an enterprise adoption programme on price alone produces a decision you will regret.

Is AI upskilling worth the investment?

It depends entirely on execution, and the evidence supports both caution and commitment. BCG reported in June 2026 that more than 60% of organisations see little to no return on AI investment and nearly 80% of AI transformations fail to deliver expected impact; the same research found leading companies invest up to twice as much as laggards in upskilling, allocate as much as 60% of AI budgets to capability building, and are four times more likely to run structured learning programmes with protected time. Run the worked example in this guide against any proposal: population, baseline, verified hours returned, value per hour. If the provider cannot fill in those four numbers with you, the investment case does not exist yet.

Sources

1. Ministry of Digital Development and Information, National AI Impact Programme: Empowering Enterprises and Workers to Transform with AI (2 March 2026). mddi.gov.sg
2. Infocomm Media Development Authority, Singapore Digital Economy Report 2025 (October 2025). imda.gov.sg
3. World Economic Forum, Future of Jobs Report 2025 (January 2025). weforum.org
4. Boston Consulting Group, AI Transformation Is a Workforce Transformation (4 February 2026). bcg.com
5. Boston Consulting Group, From AI Upskilling to AI Performance: Five Questions Every CEO Should Ask (June 2026). bcg.com
6. SkillsFuture Singapore, Committee of Supply 2026 highlights. skillsfuture.gov.sg
7. SkillsFuture and GoBusiness, SkillsFuture Enterprise Credit, including expiry on 30 November 2026 and the redesigned credit from 1 December 2026. gobusiness.gov.sg
8. Enterprise Singapore, SkillsFuture Enterprise Credit FAQ. enterprisesg.gov.sg
9. Ministry of Manpower, Appointment of Inaugural Board for the Skills and Workforce Development Agency (24 June 2026). mom.gov.sg
10. Monetary Authority of Singapore, Consultation Paper on Guidelines on Artificial Intelligence Risk Management (13 November 2025; consultation closed 31 January 2026). mas.gov.sg
11. Institute of Banking and Finance Singapore, IBF Standards Training Scheme. ibf.org.sg
12. AI Singapore, AI Apprenticeship Programme. aiap.sg
13. NTUC LearningHub, organisational profile. ntuc.org.sg