Vol. 2 [Lifestyle] ✍🏻The Art of Writing Effective AI Prompts for Your Work

When using AI, are you approaching prompts as nothing more than search terms or keywords?

To help you obtain results that truly align with your purpose, this piece introduces strategies for designing prompts with greater precision. These insights draw on perspectives from literature, anthropology, and linguistics.

1. By Culture

AI doesn't arrive neutral — it arrives shaped by the languages, institutions, and knowledge systems that dominate the internet. Between 45–60% of web content is in English, so Anglo-American rhetorical conventions end up being statistically dominant in the training data. As a result, AI may learn that Western arguments and narrative structures are the appropriate default. This authority to define what counts as the legitimate knowledge could also be explained from the lens of Bourdieu's idea of cultural capital; certain forms of knowledge carry institutional weight not necessarily because they're inherently better, but because dominant institutions treat them as the standard.

Prompt-Writing Tips for Culturally Appropriate Output

1) Name the cultural context explicitly

Don't leave it to the AI to infer. State it.

"Write a business proposal for a family-owned manufacturing firm in South Korea. The decision-maker also values long-term relationship signals in addition to short-term ROI language. Avoid strictly Western pitch deck conventions."

"Draft onboarding copy for users in South Korea. Assume a communal, relationship-first relationship with technology — not purely individual optimization."

2) Ask the AI to surface its own assumptions

After any output, follow up:

"What cultural assumptions did you apply in writing this? What might not translate outside a Western professional context?"

"Rewrite this for the South Korean audience. Identify what changed and why."

3) Have it generate multiple cultural versions

"Produce three versions of this proposal: one for a US corporate audience, one for a South Korean business context, one for a Gulf family office. Note the register differences between them."

4) Audit for absence, not just error

"What relational, communal, or trust-building elements are missing from this document for a South Korean reader?"

5) Reframe the professional narrative

"Rewrite this job description so it appeals to candidates from collectivist cultural backgrounds, where individual achievement framing may sound egotistical or even selfish."

2. By Genre / Text Type

Some may view prompting as a search query. In fact, prompting is a rhetorical act. A prompt functions as a genre, which works like a social contract with implicit rules. Different genres have different rules about speaker, audience, register, and what counts as good writing. When you write within a genre, you activate those rules. Effective communication is more likely when you follow them rather than violate them.

Every prompt has the following four components. Most people fill in one or two. The AI improvises the rest from its statistical defaults.

The more deliberately you fill in all four, the less the AI improvises — and the closer the output is to what you actually need.

Prompt-Writing Tips for Genre-Aware Output

1) Cast the speaker

"As a litigation partner preparing a client briefing for a non-lawyer CEO..."

"As a VC analyst writing an internal memo on a Series A opportunity in Southeast Asia..."

"As a brand strategist working with a heritage luxury client entering a Gen Z market..."

2) Name the output genre explicitly

Don't say "write a summary." Say what kind of summary.

"Write a one-paragraph situation-complication-resolution note in the style of a McKinsey executive briefing."

"Draft a cold email in the register of a warm referral — familiar but professional, no sales language."

"Write a memo, not a narrative explanation. Use standard memo structure: Issue, Rule, Analysis, Conclusion."

3) Give the audience, not just the topic

"Write for a skeptical CFO who has three minutes and needs to decide whether to act or shelve this."

"Write for a non-technical board member who understands business risk but not software architecture."

"Write for an academic audience familiar with qualitative methods but unfamiliar with AI."

4) Set the register in writing

"No hedging. No bullet points. Declarative sentences only. If you're uncertain, say so in one clause — don't soften everything."

"Conversational but precise. Short paragraphs. No jargon. Read like a smart newsletter, not a white paper."

5) Diagnose, don't just retry

When output misses, identify which component broke — then correct only that.

"The speaker framing is off — rewrite assuming the author is an anthropologist, not a business consultant."

"The register is too formal. Keep the content, shift the tone to a peer-to-peer conversation."

"The addressee is wrong — you're writing for the AI as a generalist. Rewrite assuming it is a senior editor at a policy journal."

3. By Narrative

AI is very good at producing information, but it is not good at producing meaning. That's why so much AI output feels flat, often forgettable.

The problem is that information and narrative are not the same thing. Narrative is what makes people read, remember, and act. AI skips it by default, so you have to make sure it is added.

Humans don't absorb facts sequentially. We need to put information in context and create meaning out of it, to comprehend and remember it.

Without deliberate prompting, AI may not include

• a person with something at stake: an actual human figure at the center of a pitch, memo, or case study that gives readers someone to identify with
• a "so what" that lands: not just conclusion, but what it means for this reader, here and now.

Prompt-Writing Tips for Human-Centered, Meaning-Making Output

1) Add a human anchor

"Add a specific person or role who is affected by this situation. Make them the entry point for the reader."

"Rewrite this case study so it opens with a person facing a real problem — not with company overview or metrics."

2) Make the "so what" explicit

"End with what this means for the reader specifically — not a summary of what was just said."

"Add a final sentence that points to the implication, not the conclusion. What does this change for someone reading it today?"

3) Force narrative structure by format

"Structure this proposal as: situation → complication → resolution. Label each section."

"Rewrite this as a problem-solution narrative. Do not use bullet points. Build toward the answer."

"Write this email so the stakes are clear in the first sentence. Background goes after."

4) Raise the stakes

"Rewrite this so it's clear why this matters — what happens if the reader ignores it, delays, or decides wrong."

"Add the cost of inaction. Not dramatically — factually."

5) By format — quick prompts

Reports:

"What is the central tension this report exists to address? Lead with that. Restructure everything else to serve it."

Presentations:

"Identify the one thing the audience needs to feel by the end. Cut any slide that doesn't build toward that."

Case studies:

"Rewrite the opening so it starts with a person and a problem, not a company description or a metric."

Emails:

"Move the stakes to sentence one. Start with the decision that needs to be made, not with context."