How to Use an AI Framework for Marketing: The EASY Method

Short answer: Work with AI in four steps: Explore, Action, Store, You. Let AI do almost all of the exploring and about three quarters of the making, but keep the checking, the gatekeeping and the decision about your positioning for yourself. Speed is now cheap; being in the right position is not.

Based on my guest talk at Lingnan University on 14 October 2026.

Sam Wong (黃力桐) is a Hong Kong corporate AI trainer and co-founder of Adaptig. 11,000+ people trained · 70+ organisations · 13 countries.

In October I went back to Lingnan University, where I studied, as a guest speaker in an e-marketing class. I gave the students one thing to remember and a four-letter way to work. This is the written version.

What do people really want from AI?

Not AI. I opened by asking the room what Hong Kong people ask me most about AI. The answer from 2025 surprised them: whether DeepSeek is accurate at fortune telling.

That tells you something. As I said in the talk: "We don't actually very much care about AI. People don't actually care about AI, even me. We care about ourselves, our fortune, our life, our career, our relationship." Underneath all of it is uncertainty.

Marketing is the same. You don't know the market, you don't know who will buy, and you don't know how your brand looks from the outside. In my words: "Marketing and business is everything about exploring uncertainties." AI is useful because it shrinks that uncertainty, not because it is AI.

Why does positioning matter more than speed?

Because everyone now has speed. I used football. The fastest strikers in the world are very fast, but "fast doesn't equal goal. You can be a fast runner, but you can be a very poor finisher." The good ones score because they are in position: right place, right time.

So the one line I asked the students to keep was: "Find your positioning in the era of AI." Producing more with AI is easy. Deciding where you want to stand, and checking that AI is moving you there, is the work.

What is the EASY framework?

EASY is the order I work in with AI, every day:

Step	What it means	AI's share	Your share
E: Explore	Go wide on a new problem: market, audience, options	about 99%	the last 1%: check it makes sense and the sources exist
A: Action	Analysis (noise into signal) and Attention (make work people notice)	about 75%	the last 25%: gatekeeping, fact-checking, fine-tuning
S: Store	Save what worked, what failed and why, in one place	does the pattern-finding	choosing what goes in
Y: You	Decide what you want to be known for	about 1%	about 99%

The percentages are my rule of thumb, not a measurement. The point is where the human part sits.

The name comes from the pitch too. In a tight moment, a teammate shouts "easy, easy". Working with AI should feel like that.

How should you explore a new market with AI?

Talk, don't type. My three-word trick is stand, walk and talk: I use voice mode and think out loud instead of sitting in front of a monitor. I wrote up the method in how to use ChatGPT at work without typing.

Two more habits:

Ask more than one AI. "Different models have different training materials, so they will give you different responses. And then it is up to you to select which one is reasonable." In the talk, on a brief for a new energy-drink brand, two voice assistants recommended two completely different partners.
Ask for one word. After the AI makes a pick, ask it to justify the pick in one word. It forces the reasoning to the core and skips the padding.

Then hold the last 1%. Like a software installer that races to 99% and then crawls, the final step takes time: read the output, open the sources it quotes, confirm they exist. "Even if you are using a paid, smarter model, you still need that 1%."

How much of your marketing work should AI do?

In making things, about three quarters. Use it to build prototypes fast, such as a landing page with a sign-up form or a first poster draft, and to argue with your plan: I ask AI to act as a sceptical marketing director and tell me where my campaign plan and survey results don't make sense.

The final 25% is yours. "You need to do gatekeeping. You need to do fact-checking. You need to do fine-tuning."

We have all seen brands publish AI work without enough human judgment and get a backlash. My interpretation is that the problem lies not with AI itself, but with how the work was produced. And the rule behind it: "If you want to get attention from somebody, but you are not willing to spend enough effort and sincerity and attention on the work, it is unfair for you to expect that attention back."

Why should you save your success stories?

Because AI is good at patterns, and patterns need material. Its strength is not creativity: "AI is good for finding patterns. The more context you give to the AI, the better its pattern-finding ability."

Many brands lose the formula: a campaign works, and a year later nobody can find why. Start simple: "build your success system by just saving everything into one folder." Notion or Obsidian can come later; a Google Drive folder is enough to begin.

Two things to save:

Other people's success. I collect strong examples, such as a famous TV pitch I treat as a textbook case in product and marketing, and put the videos and pages into NotebookLM so I can question them, turn them into audio, or build a study guide.
Your own thinking. "Save the thinking, not just the output." Export your AI conversations to PDF and keep them with the work. People love a making-of; your own chain of thought is the making-of of your work.

What does AI leave for you to do?

The part that defines you. For the last letter, I asked the students to spend almost all the effort on themselves and use AI for about 1%.

Write a positioning blueprint, the same one a brand uses:

Niche: what do I want to be known for?
Audience: who should know my name?
Pain point: what problem keeps them up at night?
Alternatives: who else is on their shortlist?
Why me: what is my unique proposition?

Save it and give it to your AI, so its answers start from who you are.

I closed with this: "It's 0% about AI: your honour, your glory, your credibility. But 100% it is your responsibility ... No one is going to blame AI for that in real life."

How can a marketing team start using EASY?

Pick one live project and run one week of it:

Explore it by voice with two different AI tools, and ask each for one-word reasons.
Prototype one asset and give the last 25% to a named person.
Start one shared folder for results, including what failed.
Write the team's five-line positioning blueprint and paste it into your AI tool's instructions.

If you are planning this for a team, my page on corporate AI training in Hong Kong explains how I run sessions. 中文讀者可看香港企業 AI 培訓。

FAQ

What is the EASY framework for AI in marketing?

EASY is a four-step order for working with AI: Explore, Action, Store, You. AI does most of the exploring and about three quarters of the making. You keep the checking, the final 25% of gatekeeping and fine-tuning, the choice of what to save, and almost all of the work on your positioning. I taught it at a Lingnan University guest talk on 14 October 2026.

How much should I rely on AI for market research?

For first exploration, heavily: I let AI do about 99% of the going wide. But hold the last 1%. Read what it gives you, open the sources it quotes and confirm they exist, even with a paid model. Ask more than one AI, because different models are trained on different material and will give different answers.

What is the one-word trick in AI prompting?

After an AI recommends something, ask it to explain its choice in one word. Forcing a single word makes the reasoning condensed and skips the padding, so you see the core rationale quickly. It works well in early exploration, when you are comparing options from several AI tools before you commit to a plan.

Why is positioning more important than AI speed?

Because everyone now has the speed. Like a football striker, a fast runner can still be a poor finisher; scoring depends on being in the right place at the right time. With AI, the advantage comes from deciding where you want to stand and checking the work moves you there, not from producing more.

How do I stop AI marketing content from backfiring?

Keep the last 25% human: gatekeeping (should this go out?), fact-checking (is every claim true?) and fine-tuning (does it sound like us?). Audiences notice when little effort went into a piece. If you want people's attention, the work needs your attention and sincerity first.

What should I save in an AI second brain?

Your work, feedback, results and what failed, plus strong examples from others and your own AI conversations exported to PDF. Start with one folder on Google Drive or your computer. AI is good at finding patterns, so the more relevant context you keep, the more useful its analysis becomes.

Talk to me

If your team wants to work this way, message me on WhatsApp: Message Sam on WhatsApp

Sources

My guest talk at Lingnan University, Hong Kong, 14 October 2026.