Sam Wong.黃力桐Let’s talk training

TRAINING DECISIONS

AI Training for Manufacturers: What Factory Heirs Need

Published:

Short answer: AI training for a manufacturer should start from what the owners are worried about, not from tools, because factory heirs have already heard plenty about AI for marketing and going overseas and what they need is to see how it lands in their own factory. That is how I designed my 10 September 2026 session for factory second-generation successors in Guangzhou: their worries first, then my framework carried by one real case.

I am Sam Wong (黃力桐), a Hong Kong corporate AI adoption trainer and Co-Founder & AI Transformation Expert at Loopem. I have trained 11,000+ people across 70+ organisations in 13 countries, and my master's degree at the Chinese University of Hong Kong focused on records management. This piece is about one session I designed for Loopem, and what I learned about designing for an audience that is not my own.

Planning AI training for a factory or family business? See corporate AI training in Hong Kong.

Who are factory second-generation successors, and what worries them?

Factory second generation (廠二代) are the sons and daughters taking over family factories. The audience I prepared for came from food manufacturing and from homeware and furniture.

When I talked the session through before the event, this is how I described them: "They're not as well off as outsiders think. The foreign-trade environment hasn't been good lately, and inside they face a lot of problems too." Many are new successors working while the father's generation is still in the factory.

That matters for training design. A room like this does not need to be convinced that AI exists. They need to know where they can start without a fight.

Why do AI talks lose factory owners?

Because most of them sound the same. Before the session I noted where the market was: "Most of the AI trainers and speakers… are focusing on AI tools, AI workflow, how to optimize their work efficiency with AI, how to do marketing with AI."

Factory owners have heard it. As I put it to myself: if I teach them AI for finding new customers, "they've probably heard plenty. I can showcase how to use it, but landing it, landing it is what matters most." I was also sure they already understood the basics of using AI to go overseas.

There is a second trap. Much of what you read about AI in manufacturing, especially online in mainland China, is about machines and big, impressive projects. That picture rarely matches where a family factory can begin.

What should AI training for a factory start with?

With their anxiety, not my framework. This was the hardest call for me, because I do have a framework and I like it.

My first instinct was to open with training and corporate memory. Then I caught it: "If I open with training, corporate information, passing on corporate memory, that may not touch their core anxiety." So I wrote down the rule I used for the whole design: "This framework is what I want to say, but the communication skill is to say first what they want to hear."

I also wanted nobody to leave thinking I had come to sell them AI.

Is recording a master craftsman's know-how with AI enough?

No, and I had to admit that to myself. One idea I had: when an experienced master trains a newcomer, record it, analyse it with AI, and learn "which worker and which step tends to go wrong." But, in my own words at the time, "that is still only a record."

The same problem showed up when I played the role of a new successor: even if AI has recorded the company's information and memory, "it still doesn't become something with influence." A record becomes useful only when it changes a decision or a daily routine. That gap, from record to influence, is what a factory session has to close.

What is the Archive, Algorithm, Action roadmap?

It is the three-step route I use for AI adoption, built on my records-management background. Everyone talks about tools, and about forward-deployed engineering, but "there's no workable roadmap for how those tools land."

For factory heirs, I decided the roadmap should not appear as a model on a slide. It should arrive inside a story.

What case should an AI session for manufacturers use?

One case, told from start to finish. I planned to build the session around a Hong Kong food manufacturer I have advised, whose AI transformation follows the same route. The aim: "one through-line across the whole session, a case they can understand, showing my framework from start to finish."

I will be honest: this is the part I find hardest. "I often don't know how to do case sharing, because I don't know which cases they'd be interested in, or what I need to say to show them why the case matters." For the furniture and homeware owners in the room, I still had to work out how the same story would translate.

How I prepared: thinking out loud with a voice AI

I prepared both sessions by talking to a voice AI and recording it. Four moves kept it from simply agreeing with me:

  1. A role. "I'm the factory second-gen, you're my chief of staff."
  2. A ban on agreeing. "I need you to be direct with me. No need to agree with me."
  3. Pushback on its own advice. "How can I be sure it works?"
  4. The audience's motive. "When they sit down at this session, what do they want to hear?"

The AI's answers were useful, but most of the design above came from my own questions and worries. It surfaced them because I asked it not to be polite.

FAQ

What should AI training for manufacturers start with?

It should start with the owners' real worries, not with AI tools. In my 10 September 2026 session design for factory heirs in Guangzhou, I opened from their situation: a hard foreign-trade environment, internal problems, and a father's generation still in the factory. A framework or tool demo comes after the room feels understood, and only to answer that worry.

Why do factory owners lose interest in AI talks?

Most AI talks repeat the same tool layer: efficiency, marketing, finding new customers and going overseas. Factory owners have heard that many times. What they have not seen is how AI lands in their own factory. A session that only showcases tools sounds like a sales pitch. A session about landing one change keeps them listening.

Is recording expert know-how with AI enough to change a factory?

No. Recording a master craftsman training a newcomer and analysing it with AI can show which worker and which step tends to go wrong. But on its own it is still only a record. It creates value only when it changes a decision or a daily routine, such as how new staff are trained or checked.

What is the Archive, Algorithm, Action roadmap?

It is Sam Wong's three-step route for AI adoption, built on his records-management background. Archive puts the company's memory and knowledge in order. Algorithm builds rules and a controllable environment around what is unstable. Action turns daily operations into a closed loop that keeps learning from the work. In a factory session it works best inside one real case, not as a slide.

How did Sam Wong prepare the factory second-generation session?

He talked the design through out loud with a voice AI in two recorded sessions on 2 and 6 September 2026. He gave the AI a role ("you're my chief of staff"), told it not to just agree with him, pushed back on its advice, and kept asking what the audience wanted to hear. The final structure came from his own questions and worries.

Talk it through

If you run a factory or family business and are thinking about AI training for your team, I am happy to hear where you are first. Message me on WhatsApp, or read more about corporate AI training in Hong Kong.

Sources