At the Most Critical Moments at Work, You Usually Have No Phone in Your Hand
Illustration · not a service recordAt the most critical moments at work, you usually do not have a phone in your hand.
Anyone who has seen Mission: Impossible or a 007 film knows this: however thrilling the chases by plane, sports car or motorcycle, what decides the outcome in the end is often the hero getting out of the vehicle and running, climbing, or standing in front of the other person and talking.
This is my slide from today’s corporate AI workshop at the Singapore Business Federation (SBF), where the audience was HR people, operations managers and business owners. If it makes you think of spy movies, that is exactly right. The first four are AI: the plane is the most powerful model, the F1 car is specialized AI, the car is the everyday tool, and the motorbike is the little helper. The last two are you, and the slide says OR NOT: the bicycle is you plus simple tools, and the last one is you, running on your own.
At work, when are the moments for the bicycle and your own two legs?
My answer: the moments when you are handling trust.
Telling a client, “This time it was our fault.”
The last five minutes at the negotiating table.
Keeping a colleague who already wants to leave.
Sitting down with a parent to talk about what has been going on with their child lately.
Delivering a piece of bad news at a bedside.
These moments call for sincere conversation, and chances are you will not have a phone in your hand. You can hardly open up an AI right in front of the other person and ask, “What should I say now?”
So you have to train your legs in ordinary times. The brain is like a muscle: it only grows when you use it. If you ride for every stretch of the road, your legs will give way at the critical moment.
This is also why Junyi Academy’s AI Notebook deliberately does not give the answer first. In New Taipei, 227 sixth-graders used it for a full school year, and questions that asked directly for the answer fell from 22.6% to 5.6%.
The workplace is the same. New hires used to follow their managers to meet clients and take customer complaints; that was leg training. If all of this is handed over to AI, who will handle trust five years from now?
I am certainly not saying do not use AI. Junyi has about 50 people, and we use several kinds of AI every day. I recently tried using AI to design my writing in more depth, and in three weeks my social media views went up 26-fold.
But what I want to say, and what I think matters more, is this: the time AI saves should be spent on the moments without a phone. AI can take you as far as the door; pushing it open, sitting down, and looking at the other person as you speak is something you have to do yourself.
Otherwise, you could also:
ask AI to apologize for you, and expect the other person to forgive you;
ask AI to meet clients for you, and expect them to trust you;
ask AI to keep your child company for you, and expect that child to tell you everything when he grows up.
Junyi’s AI literacy framework, “Ask, Use, Manage, Build (問用管造),” works the same way: Ask and Use are things one person can start alone; Manage and Build need a group of people who trust one another. Our partners in Singapore call it LEAD.
Ask: Learn to Ask
Use: Execute with AI
Manage: Align & Assure
Build: Design for scale
AI is the multiplier; trust is the number being multiplied. Multiply by zero and you still get zero. AI will run faster and faster, but those last few steps are still ours to walk.
My most recent critical moment “without a phone” was, in fact, today’s talk.
- Original: Ray Lu’s personal Facebook post (published 2026-10-26)
Course titles follow the source. Check each course page for its teaching language.
Generative AI Leader Certification Learning Path
Five preparation courses are free and address business adoption and transformation. Course completion is not certification; the separate exam costs US$99 plus applicable tax. Its 90-minute duration is exam time, not course length.
AI capabilities and limitations
Understand next-token prediction, knowledge, working memory, and steerability to diagnose hallucinations and context limits. No technical prerequisites.