Why is HR responsible for AI adoption rather than IT?

IT provides access; HR changes behaviour. Adoption is decided by whether people know the rules, have the skills and feel confident enough to use AI daily, which are all people questions. In Malaysia and Singapore HR also controls the training funding routes, so the programme naturally sits with HR while IT owns tool security and access.

What should an HR leader do first on AI adoption?

Run a short anonymous survey on current use and confidence, then write a one-page usage policy. The survey gives you a baseline and usually reveals far more use than expected. The policy removes the fear that stops careful employees from starting, and it protects the company from data leaks that are probably already happening.

How do I get managers to support AI adoption?

Ask them to use AI visibly on their own work and to raise it in one-to-ones. Microsoft's 2026 Work Trend Index found only 32% of Malaysian AI users see their leadership clearly and consistently aligned on AI. Managers who model use and talk about failures create permission faster than any memo from HR.

Is AI training for HR teams HRD Corp claimable in Malaysia?

Yes, when delivered by an HRD Corp Accredited Trainer through a registered training provider with the grant approved before training. In-house course fees are claimable up to RM10,500 per group per full day under the Allowable Cost Matrix. Apply at least 30 days ahead, and note that training must now start 14 to 90 days after approval.

What funding is available for AI training in Singapore?

According to EnterpriseSG, the SkillsFuture Enterprise Credit gives eligible employers a one-off S$10,000 credit covering up to 90% of out-of-pocket costs on supported programmes, with the current credit expiring on 30 November 2026 and a redesigned version announced from December 2026. Check whether a specific course is SkillsFuture-supported before counting on subsidies.

How long does it take to see results from AI adoption?

Confidence moves on the training day itself. Weekly usage settles within three to four weeks if managers follow up. Time savings on named tasks are measurable at 60 to 90 days. A 90-day cycle with one team is enough to produce a result you can show leadership and use to justify the next groups.

Should HR use AI in recruitment and performance reviews?

Use it for drafting, summarising and preparation, and keep the decision with a human every time. Your usage policy should say this explicitly. AI can draft a job description or summarise interview notes well. It should not rank candidates or write a performance rating unsupervised, both for fairness and for the regulatory direction in both countries.

What does the AI Confidence workshop cover for HR teams?

One day, seven training hours, up to twenty participants, in person at your workplace. HR teams work on their own real tasks, such as policy drafting, job descriptions, onboarding materials and employee communications, using the 4D Framework. Everyone leaves with a working AI assistant set up on their actual job. Pricing is per group and depends on size and needs.

AI Adoption for HR Leaders: A Practical Guide

Why HR ends up owning AI adoption

AI adoption arrives in HR's inbox for a simple reason: it is a people change disguised as a technology change. IT can switch on Copilot in a week. Getting 200 people to use it well, safely and consistently is a capability and culture problem, and that is HR's territory.

The data shows how far ahead employees already are. EY's Malaysia findings from its 2025 Work Reimagined Survey, published in March 2026, found 93% of Malaysian employees use generative AI at work, while only 12% receive sufficient training to benefit fully. Your people are not waiting for a policy. They are using AI on their phones and hoping nobody asks.

The second reason is money. In Malaysia the HRD Corp levy sits with HR. In Singapore the SkillsFuture schemes sit with HR. Whoever controls the training budget ends up controlling the adoption programme, whether they planned to or not.

The third reason is risk. Confidential data in public chatbots, AI-written performance reviews, and hiring decisions influenced by tools nobody validated are all HR problems the moment they surface. Owning adoption early is cheaper than cleaning up later.

The three jobs HR has

Strip the topic down and HR has three jobs. They are best done in this order, because each one makes the next easier.

01Policy and guardrails. Decide which tools are approved, what data never goes into them, who can use AI for what, and how AI-assisted work is labelled. Put it on one page and get every employee to read it. This removes the fear that keeps careful people from starting.
02Capability building. Train people on their own work, not on a product. The test of good training is whether someone leaves with a task they now do differently, not whether they can recite features. Mixed-experience groups work well if the method is structured.
03Culture and confidence. Confidence is the variable that predicts daily use. Managers who use AI openly, share what failed, and ask about it in one-to-ones create permission. HR's job is to make that behaviour visible and expected.

Most organisations start with job two, buy a course, and wonder why nothing changed. Without guardrails people hold back. Without culture the skills fade in a fortnight.

A 90-day plan for an HR leader

Ninety days is long enough to change behaviour in one part of the organisation and short enough to hold attention. Here is a sequence that works.

01Days 1 to 10: find out what is already happening. Run a short anonymous survey on which tools people use, for what, and how confident they feel on a five-point scale. This is your baseline.
02Days 5 to 15: draft the usage policy. Use the checklist in the next section. Get legal and IT to review it in one meeting, not a chain of emails.
03Days 10 to 20: choose the first group. Pick a team with recurring, document-heavy work and a manager who is curious. HR, finance and marketing are usual candidates.
04Days 15 to 30: sort out funding and dates. In Malaysia, submit the HRD Corp grant application at least 30 days before training; since 15 June 2026 training must start 14 to 90 days after approval. In Singapore, check which SkillsFuture support your company can access.
05Days 30 to 45: brief the trainer on real work. Collect the documents, templates and recurring tasks the group actually handles. The training should be built on them.
06Days 45 to 50: run the workshop. One full day, up to twenty people, everyone leaving with a working assistant set up on their own job.
07Days 50 to 75: follow up weekly. A 20-minute team stand-up where people show what they built. The manager asks about AI in every one-to-one.
08Days 75 to 90: measure and report. Repeat the baseline survey, collect time-saved examples from three named tasks, and present a one-page result to leadership with a recommendation for the next two groups.

What to put in an AI usage policy

A usable AI policy fits on one page. Anything longer will not be read. Cover these points and stop.

Approved tools, by name, and how to request a new one.
Data that must never be entered into any AI tool: customer personal data, salary data, unreleased financials, anything under NDA.
Where AI may be used freely (drafting, summarising, research) and where a human must own the decision (hiring, performance, disciplinary, legal, medical).
How to check AI output before it leaves the building, and who is accountable for it. The person who sends it, always.
Whether and how to label AI-assisted work for clients or regulators.
What to do when something goes wrong, and a named person to ask.
A review date. Tools change every quarter; the policy should be looked at twice a year.

Write it in plain language and in the languages your workforce reads. A policy people understand is a policy people follow.

How to choose training that changes behaviour

Most AI training teaches tools. Tools change every quarter, so that training expires quickly. What lasts is a way of thinking about when to delegate to AI, how to describe what you need, how to judge the result, and how to follow through. That is what to buy.

Use these five tests when you compare providers.

01Do participants work on their own documents and tasks during the session, or on the trainer's examples?
02Does everyone leave with something working, such as an assistant set up on a recurring task?
03Can the provider show survey evidence from past groups, not just client logos?
04Can the trainer handle a mixed group, from sceptics to power users, without losing either end?
05Is the trainer accredited for the funding route you need: an HRD Corp Accredited Trainer in Malaysia, or a provider recognised under the relevant SkillsFuture scheme in Singapore?

The AI Confidence workshop is built to pass all five. It uses the 4D Framework (Delegation, Description, Discernment, Diligence), works on participants' real tasks, and every participant leaves with their own working AI assistant. It runs in one day, or two days with more hands-on labs, for groups of up to twenty.

Measuring adoption: usage, confidence, time saved

You cannot show the board a feeling. Measure three things, before and after, and adoption becomes a number.

Usage. The share of the group using AI at least weekly on work tasks. Ask, do not rely on licence dashboards, because much of the real use happens on personal accounts.
Confidence. Self-rated on a five-point scale. This is the leading indicator: confidence rises before usage does, and it predicts who will keep going after the trainer leaves.
Time saved. Pick two or three named recurring tasks and ask people to estimate hours per week before and after. Keep it rough and honest.

As an example of what these measures can look like, the AI Confidence workshop's own survey data from 100 verified responses across hospitality, tech and manufacturing shows confidence using AI rising 42%, from 3.0 to 4.3 out of 5. In the same data, 90% called it a worthwhile use of their time and 84% left ready to apply what they learned immediately. Whatever programme you choose, ask for its equivalent numbers.

The Malaysia and Singapore specifics

Malaysia. If your company is registered with HRD Corp and pays the levy, in-house AI training by an HRD Corp Accredited Trainer is claimable up to RM10,500 per group per full training day under HRD Corp's Allowable Cost Matrix (January 2026 version). Claims run through a registered training provider. Sylvia is an HRD Corp Accredited Trainer and liaises with her provider partners on your behalf, so you are not managing the provider side yourself.

Timing matters more since HRD Corp's Employer's Circular No. 2/2026 took effect on 15 June 2026. Training must start between 14 and 90 calendar days after grant approval, and approved grants cannot be amended. Fix your dates before you apply.

Singapore. There is no HRD Corp equivalent, but support exists. According to EnterpriseSG, the SkillsFuture Enterprise Credit (SFEC) gives eligible employers a one-off S$10,000 credit that supports up to 90% of out-of-pocket expenses on supported programmes, with the current tranche expiring on 30 November 2026. SkillsFuture Singapore has announced a redesigned SFEC from December 2026. Course-level subsidies apply only to SkillsFuture-supported courses, so check a programme's status before you assume funding.

The AI Confidence workshop is delivered in person anywhere in Malaysia and in Singapore. HRD Corp claims apply to Malaysia only.

How Should HR Lead AI Adoption?