AI isn’t magic. It’s a skill stack. After watching Jeff Su break down how pros use AI, I’m convinced far too many teams are stuck tinkering at the edges. My take is simple: treat AI like a colleague or get left behind. This matters for anyone who writes, sells, builds, or leads—because the winners aren’t the ones who try the most tools. They’re the ones who build repeatable systems.
The Core Argument
Jeff Su argues that professionals fall into three camps: curious, literate, and native. The leap from “literate” to “native” isn’t about new apps. It’s about new habits. Redesign your workflow as if an AI collaborator already exists.
“Organize your information by where you will use it, not where you found it.” — Jeff Su
That line hit me. I spend my life helping brands turn customers into superfans, and the secret is the same: make it easy to repeat what works. AI should be a system, not a stunt.
What Jeff Gets Right
Three habits stood out as practical, repeatable, and high leverage:
- Leave AI breadcrumbs: Link your best AI chats directly inside the docs where you use the output. Add context next to each link.
- Build an AI swipe file: Save great proposals, decks, emails, and reports. Ask AI to analyze patterns, then apply them to your task.
- Plan tasks AI-first: Break big projects into microtasks. Tag which ones are manual and which ones AI should handle, and choose the right model for each.
Each habit reduces friction and raises the floor on quality. Most people know AI can help. Few decide where and how it will help before they start.
“Most people should just pick one AI chatbot and get really good at it.” — Jeff Su
That’s smart. Tool-hopping kills momentum. Pick a model, master it, and document what works.
Evidence That This Works
Jeff shows how “breadcrumbs” speed recall: link the chat where you refine a prompt, brainstorm an outline, or apply storytelling—right inside the working doc. No more digging through old threads. I’ve seen the same with marketing teams. The moment they anchor AI chats to live projects, drafts improve and handoffs get cleaner.
“Start narrow and expand gradually.” — Jeff Su
With swipe files, he uses top-tier examples (think consulting decks) to train the model on structure and tone. That’s not cheating; it’s coaching. You’re showing AI what “good” looks like. The output gets sharper, faster.
On planning, his example of a weekly product newsletter is on point. He breaks the work into steps, uses Notebook LM to check facts, then switches to a creative model for voice. Match the task to the model. That’s the leap from dabbling to doing.
My Playbook for Becoming AI Native
I agree with Jeff’s framework, and I’d push it further with a “fan-first” lens—because loyalty is built in the details your competitors skip.
- Make a reusable “AI cover sheet.” For any repeat task, keep a one-pager with goal, audience, tone, do/don’t style notes, and success metrics. Paste it at the top of every prompt.
- Rate outputs like a coach. After each session, add a one-line verdict and what you’d change next time. Link it in your breadcrumb.
- Run a weekly 20-minute “prompt retro.” Save the best prompts to a shared library by use case. Kill the ones that waste time.
These steps turn scattered wins into a playbook your team can use every week.
Answering the Doubters
Some worry that AI will flatten voice or copy errors. That happens when teams skip process. If you plan AI use up front, keep a swipe file of gold-standard examples, and fact check with the right tool, quality goes up, not down. And if you’re leading a brand, AI doesn’t replace your point of view—it amplifies it.
What You Should Do Next
Stop treating AI like a toy. Treat it like a teammate with a job description. Here’s a quick start for your next big project:
- Spend five minutes listing microtasks and tagging AI/manual.
- Pick one model and stick with it for a month.
- Create a swipe folder for one use case you do weekly.
- Link your best chats in the working doc with one-line context.
Jeff Su is right: most professionals stall at “literate.” The edge goes to the ones who systemize. If you care about winning hearts and wallet share, your process has to scale. Your customers won’t remember how many tools you tried. They will remember how fast and how well you served them.
The choice is simple. Keep dabbling and hope for sparks—or build the machine that makes sparks on command. Start today.
