HRDF Field Notes — 2026-05-28

This Thursday Field Notes edition tracks three patterns we are seeing across Malaysian L&D teams this fortnight: SBL-Khas pre-approval cycles stretching longer in the wake of the HRD Corp governance reset, AI cohorts moving to blended online-plus-classroom formats to fit per-day claim caps, and an early uptick in employer interest in agentic-AI course codes now registered under HRD Corp.

What's new

• Governance reset still unfolding. Free Malaysia Today reported on 21 February 2026 that three more HRD Corp officials were suspended, following the broader governance reset announced earlier the same week. The official HRD Corp media release of 21 February 2026 framed this as a step-up in governance and compliance discipline. Operationally, processing queues for SBL-Khas pre-approvals have stretched longer than 2025 norms — most visibly for course codes that are new to an employer's claim history.
• 2026 trainer fee bands and per-day caps are confirmed. Per guidance collated by QuickHR, CorporateTrainingMY, and Pertama Partners, the current allowable trainer fee range under SBL-Khas remains RM80 to RM200 per hour depending on certification level. Group-training caps hold at RM10,000 per day per trainer and RM1,750 per day per participant for individual training. SMEs under 200 staff continue to access 100 per cent SBL-Khas subsidy; other employers sit in the 50 to 80 per cent band depending on training type.
• LMS and online AI training remain first-class claimable formats. Disprz and Pertama Partners both note that in 2026 HRD Corp recognises LMS subscriptions, virtual training, and analytics-driven learning programmes. Online-only AI training is claimable up to RM300 per participant per programme provided the provider is HRDF-registered, usage is tracked, assessments exist, and certificates are issued.
• Agentic-AI course codes are gaining traction. Trainocate Malaysia has listed its Applied Skills training including Building an Agent-First App as HRD Corp claimable, alongside the AB-100 Agentic AI Business Solutions Architect certification path. This is a meaningful shift: the claimable register is starting to accept agent-framed training rather than only generative AI productivity courses.

Why it matters for your training claims

The governance reset is the dominant operational signal this quarter. When HRD Corp tightens compliance reviews, the visible effect at employer level is longer wait times between SBL-Khas application submission and approval. Teams who were used to four-to-seven working day turnarounds on familiar course codes should plan for ten working days through May and June, and longer for any course code that is new to your account history. Retroactive claims remain unavailable, so the real cost of an under-planned timeline is a cancelled cohort or an out-of-pocket training spend that should have been claimable.

The trainer fee bands matter because the RM80 to RM200 per hour range sits below the prevailing market rate for senior AI practitioners running technical sessions. Employers who want the top tier of trainer for high-impact AI programmes are increasingly structuring engagements as a blend: claimable SBL-Khas hours at the certified-trainer rate, supplemented by paid consultancy hours outside the levy. This is legitimate when the training and consultancy components are clearly delineated and documented, but it requires more attention to scope-of-work and invoicing than the simple full-claim model that many L&D teams have defaulted to in recent years.

The agentic-AI register-eligibility shift is the most interesting medium-term signal. Through 2024 and most of 2025, HRD Corp claimable AI training was effectively confined to productivity-tool literacy: ChatGPT, Copilot, prompt engineering, AI for office work. Courses that taught how to design and operate AI agents — orchestration, tool use, evaluation, agent-based workflow design — were harder to fit under existing course codes. The recent acceptance of agent-framed Microsoft and applied-skills training opens room for more sophisticated cohorts under SBL-Khas. We expect more providers to file under this category through H2 2026.

What to do next

1. Add five working days of buffer to every SBL-Khas pre-approval timeline you scope between now and the end of Q3 2026. If your training dates are fixed, the application must land earlier — not closer — than your historical norm.
2. Review your training provider's trainer-fee structure for any AI cohort. If senior practitioner hours are required, ask the provider to model both a full-claim and a blended (claimable training plus non-claimable consultancy) structure so the internal budget conversation has two defensible options.
3. Map your H2 2026 AI training needs against the agentic-AI course codes now available under the claimable register. If your roadmap includes AI agent build-out, multi-agent workflow design, or AI risk and governance, those topics are increasingly fundable under SBL-Khas rather than purely budget-funded.

Based in Penang — delivering nationwide. Contact [email protected] to scope your next HRDF-claimable cohort.