Training budgets are getting larger. Companies across Hong Kong are investing in AI upskilling programs, certification courses, and workshop series. But most are training without a baseline. They don't know what their teams know before training starts, so they can't measure whether the training actually worked.
This is like medicating without diagnosis. You might help. You might hurt. You almost certainly won't hit the target.
The Unmeasured Training Problem
Without baseline assessment, several bad things happen:
Repetitive Training — You send your product manager to an "AI Fundamentals" course. But they already know fundamentals. Three days wasted. Meanwhile, they actually need to understand model inference, which nobody trained them on.
Wrong-Level Training — You bring in a general AI course for your entire department. Senior engineers are bored. Junior engineers are lost. Everyone in the middle gets partial value. A one-size-fits-all approach fits nobody.
Untracked ROI — You spend HK$200K on training. Did it improve decision-making? Did it accelerate projects? You don't know because you didn't measure before and after.
Wasted Talent — You train your entire team, including people who already have expertise. You've just paid to re-teach people things they already know, while neglecting the gaps that actually matter.
Why Companies Train Blind
Most organizations have never established baseline knowledge. They don't have clear benchmarks for what a Product Manager should know vs. an Engineer vs. a Finance Manager. So they default to generic training for everyone.
They also assume training needs are obvious. "We need AI skills, so let's do AI training." But that's not how skill development works. What your Finance team needs to know about AI is completely different from what your Engineering team needs. And both are different from what your Business Development team needs.
Without measurement, you're essentially guessing.
The Measure-First Approach
Companies that get real ROI from training measure first. Here's what they do:
1. Baseline Assessment — Measure what your team knows before any training. This shows you the starting point and reveals patterns (e.g., "finance managers consistently weak on model validation").
2. Role-Specific Benchmarks — Define what success looks like for each role. A Product Manager's AI knowledge should include understanding feasibility and project timelines. An Engineer's should include model training and deployment. A Finance Manager's should include ROI evaluation and risk. These are different targets.
3. Gap Identification — Compare actual knowledge against role benchmarks. Now you know exactly what training will move the needle. Instead of "AI skills training," you design "Model Validation for Finance Managers" or "AI Project Feasibility for Product Owners."
4. Targeted Training — Deliver training that addresses specific gaps. This is 3-4x more effective than generic training because it's designed to fix actual problems.
5. Post-Training Measurement — Measure knowledge again after training. This tells you what worked, what didn't, and what still needs attention.
The ROI Math
Let's say your training budget is HK$300K for the year:
- Unmeasured approach: 40% of budget goes to redundant training, 30% goes to wrong-level training, 30% actually adds value. Net: HK$90K in real value.
- Measure-first approach: 95% of budget targets actual gaps. Net: HK$285K in real value.
The same budget, 3x better outcomes, because you measured before spending.
What Gets Measured Gets Done
There's a reason the saying exists. When you measure AI knowledge before training, you create accountability. Teams know they'll be tested afterward. Trainers know they're being evaluated. And you get honest data about what actually improved.
More importantly, measurement creates focus. Instead of broad AI literacy courses, you run targeted programs. Instead of hoping training sticks, you track results. Instead of guessing whether you're competitive, you have data.
Starting Your Measurement Program
Take the AI IQ Test with your team. Get a baseline. Identify your biggest gaps. Then design training that actually fixes them.
If you want to build a measurement framework for your entire organization, reach out at sales@ud.hk or WhatsApp (852) 9696 7545. We'll help you set benchmarks, identify gaps, track training ROI, and build a sustainable AI upskilling program.
Your competitors are measuring. The question is: how much longer will you train blind?