The stat is attention-grabbing, but it's also misleading. When researchers ask "Do you lack AI skills?", 52% of Hong Kong companies say yes. But the real problem isn't the lack of skills—it's the lack of clarity about which specific skills are missing.
Most companies know something is wrong, but they can't pinpoint what. That fog is expensive.
The 52% Problem Unpacked
The 52% figure comes from multiple sources, including Deloitte and PwC surveys. But it hides important nuance. Some of that 52% lack foundational AI knowledge. Others have decent AI knowledge but lack the skills to execute at scale. Still others have plenty of knowledge but lack confidence in specific domains (data governance, bias detection, etc.).
The problem: most companies don't know which category they fall into. So their training and hiring efforts are scattershot. They bring in generalists when they need specialists. They train junior staff when they need to upskill management. They build teams that look good on paper but can't execute.
What the Missing 48% Aren't Saying
Here's the uncomfortable part: 48% of Hong Kong companies didn't admit to lacking AI skills. Does that mean they have them? Maybe. Or it could mean they don't realize they're missing something. Confidence bias is real, and AI literacy gaps are especially vulnerable to it.
We often find that companies reporting "we have AI skills" actually have:
- One person who knows AI (and is probably overworked)
- Inherited knowledge from someone who left the company
- Basic training from 2-3 years ago that's now outdated
- Theoretical knowledge that doesn't translate to execution
- Know-how in one area (e.g., model training) but blind spots in others (e.g., model deployment)
That's not having AI skills. That's having a few pieces and assuming you have the whole puzzle.
The Cost of Vague Gaps
When companies know something is missing but can't identify what, they tend to overspend on training and underspend on strategic hiring. They run generic "AI literacy" courses when they actually need specialized role-based training. They hire "AI experts" without specifying what expertise they need.
The result: budget wasted on training nobody needs, talent hired that doesn't fit, and gaps that persist because they were never actually identified.
The Specificity Problem
"AI skills" is too broad. Better questions are:
- Can your team explain how a Large Language Model actually works?
- Can your data team handle training, validation, and test sets properly?
- Can your product team assess whether an AI feature is worth building?
- Can your leadership team ask intelligent questions about AI roadmaps?
- Can your compliance team identify AI-related risks in your industry?
These are specific skills. And the odds that your company has all of them? Probably low. The odds that your company knows exactly which ones you're missing? Even lower.
Role-Based Gaps Are the Real Problem
AI skills aren't monolithic. A data engineer needs different AI knowledge than a product manager, who needs different knowledge than a finance manager. Generic training treats everyone the same. That wastes time for people who already know the material and leaves others behind.
The companies making smart AI progress aren't the ones saying "everyone needs AI literacy." They're the ones saying "Finance managers need to understand how AI affects our models. Product managers need to understand feasibility and ROI. Data engineers need to understand the full ML pipeline. Let's measure what each group actually knows and train from there."
How to Identify Your Specific Gaps
You need three pieces of data:
1. Individual Assessments — What does each person actually know? The AI IQ Test captures this across your organization, showing you knowledge distribution by role, department, and seniority.
2. Role-Based Benchmarks — What should your Finance Manager know about AI vs. what should your Product Manager know? These benchmarks differ significantly, and most organizations don't have them.
3. Gap Analysis — Once you know what people know and what they should know, you can identify the precise gaps. Not "we lack AI skills." But "our product managers don't understand model training complexity" or "our finance team can't evaluate AI ROI" or "our data team lacks governance knowledge."
Once you have that specificity, your training spend multiplies in effectiveness. You hire smarter. You train smarter. You make better strategic decisions.
The Test That Reveals Your Gaps
The AI IQ Test is designed to do exactly this. It shows you:
- What your team knows about AI fundamentals
- Which roles have the strongest and weakest AI knowledge
- Which concepts are well understood and which need training
- How your organization compares to Hong Kong benchmarks
Instead of a vague 52% problem, you get specific data. Instead of general "AI training," you can build targeted programs. Instead of hoping your team has the skills, you know exactly what you have and what you need.
Next Step
Take the AI IQ Test and stop guessing. Get specific data about your team's AI knowledge. Then share it with your L&D team, your hiring managers, and your leadership. Use it to build a real AI skill strategy.
If you want to benchmark your entire organization and get specific recommendations, contact us at sales@ud.hk or WhatsApp (852) 9696 7545. We'll help you turn that 52% into actionable intelligence.
The companies ahead of you aren't guessing anymore. Are you?