Let’s be honest: saying you “use AI” isn’t special anymore. It’s the price of admission. After watching Jeff Su lay out four skills to rise above checkbox competence, I came away with a clear conviction: the winners won’t be the people who type the fastest into a chatbot, but those who design smarter systems, tell sharper stories, and keep their thinking strong. That’s not just good advice for productivity geeks. It’s a roadmap for anyone who wants fans, customers, and teams to actually care.
The Big Idea: Mode-Select Like a Pilot
Jeff’s core message is simple and right: your edge is picking the right “mode” for every task. Sometimes you hand work to AI. Sometimes you co-create. Sometimes you keep your hands on the wheel.
“Being good at ChatGPT is now the bare minimum… The cockpit rule is a mental model for deciding when to delegate, collaborate, or avoid AI.”
That “cockpit” framing sticks because it’s practical. He pairs it with Ethan Mollick’s cost–benefit lens: estimate your own time, AI’s odds of success, and the oversight time. As a leader, I’ve seen careers move faster when people stop guessing and start scoring tasks this way.
Build The Rails, Not Just The Train
Jeff argues that process design beats raw output. I agree. In my world of customer experience, the best brands don’t just answer faster; they answer the same way, the right way, every time. That only happens when you set rails the system can glide over.
“Your competitive advantage is no longer doing the work. It’s designing the process so AI can do it for you.”
He cites a telling stat: a single prompt generated working code 48% of the time; a structured workflow using the same model hit 95%. That’s not magic. That’s operations. And a Harvard–BCG study found consultants using clear handoffs with AI outperformed those who poked at tools with no plan. Process is the multiplier.
- Map a recurring deliverable into steps.
- Mark each step as autopilot, collaboration, or manual.
- Redesign the autopilot steps first for quick wins.
Once the rails are set, you get speed without chaos—and results you can trust.
Storytelling Beats Data Dumps
As someone who helps brands create superfans, this point made me nod hard: information is cheap; meaning is rare. Jeff nails it with a blunt truth—AI can list facts, but it can’t care. People can.
“If you can turn data into a story that moves people, you’re safe. If you just pass along the data, you’re replaceable.”
His practical picks—the ABT (And–But–Therefore) and SCQA frameworks—force a clear arc: set the scene, add conflict, resolve. I’ve watched executives win budgets with weaker data but stronger stories. That’s not spin; that’s understanding how humans decide.
- ABT: And = status, But = tension, Therefore = action.
- SCQA: Situation, Complication, Question, Answer.
Use either to turn dry updates into decisions that stick.
Protect Your Mind With Manual Override
Here’s the take many miss: overusing AI dulls your edge. Not your brain itself, but your habits. Jeff cites research showing that workers who let AI think first stop checking assumptions and miss edge cases. Radiologists who formed their own view before consulting AI stayed accurate. That tracks with what I see on teams: independence first, tool second.
“Think first, prompt second… When AI gives you an answer, ask: How would I verify this? What’s the counterargument?”
I practice this in my own work. I sketch the “so what” before I ask for a summary. Then I interrogate the output. That small delay keeps judgment sharp and prevents lazy agreement with a slick answer.
Where I Agree—and What I’d Add
Jeff’s framing is sharp and useful. I’d add one more filter for anyone building a fan-first brand: measure audience impact, not just task speed. Faster isn’t better if it dulls voice or trust. Use AI to scale the right moments, not every moment.
- Protect voice: lock tone, phrases, and no-go lines in your prompts.
- Stress-test for edge cases: run “what if” drills before launch.
- Close the loop: measure clicks, replies, and sentiment—not only output volume.
Speed wins the inbox. Meaning wins the relationship.
My Challenge To You
Stop bragging about using AI. Prove you can choose the right mode, build the rails, and tell a story that moves people to act. Keep your thinking alive.
Try this this week:
- Pick one workflow and score each step: autopilot, collaboration, or manual.
- Rewrite one update using ABT or SCQA.
- For one hard task, think first for five minutes, then prompt—and challenge the output.
The tools are here. Differentiation is a decision. Make yours.
