← All workflows
Content Workflow9 min readUpdated

The Master Brief Pipeline: 6 Assets, 2 Languages, 2 Hours

One brief becomes six EN + BM campaign assets in under 2 hours, with 35–45 min of human time. The full workflow: brief, voice doc, edit, adapt, native pass.

workflow 02 master brief

One master brief that becomes six campaign assets in English and Bahasa Melayu just within 2 hours, with 35-45 minutes of human time. 

You want to build a master brief workflow that is controllable, adheres to your brand voice document, and corrects all errors with minimal human edits. If you want to work on it manually, it will take at least two days.

This is the master workflow you will reuse with other workflows within Naven Pillai's AI Marketing workflows.

Why does this workflow matter

If you're a Malaysian marketer, you already know that campaigns in Malaysia don't run in just one language. And I'm aware that not every marketing team has the resources to work on the other languages.

As a result, the English content is written properly. In contrast, the BM content is either outsourced (which is slow and may not align with your brand voice), rushed internally, or run through an expensive translator (grammatically correct but with a high chance of sounding awkward).  

Multiply that by six asset formats per campaign. When you do the math, you'll know how fast it gets ugly.

Imagine the time and cost involved.

And that's why SMEs create fewer content assets, in fewer languages, and are unable to reach their target audience.

The fix is not about adding more members to your marketing team or hiring a better agency.

It comes down to sequencing.

Draft once against a controlled brief, edit it once, then adapt. Do not assume your content platforms; be specific. 

AI will assist you in what it is good at: drafting and creating matched content depending on your asset description.

And the human intervention happens on exactly two steps that can never be automated:

  • Judgment on the English, whether it follows your brand voice 
  • A native language expert's evaluation of Bahasa Melayu

The Workflow

Master brief (10 minutes, human) 
→ English drafts, all 6 assets (15 minutes, AI) 
→ Human edit pass (20–25 minutes, human) 
→ BM adaptation — not translation (15 minutes, AI) 
→ Native speaker pass (5–10 minutes, human) 
→ 6 assets × 2 languages

Two rules hold this workflow together. 

This workflow collapses into the ad hoc prompting everyone else is doing if you break them down.

Rule 1: The AI (LLMs) never invents facts. Every claim it may make lives in the brief. Every claim it may not make lives there too.

Rule 2: Bahasa Malaysia is adapted, never translated. Many experts have claimed that generative AI tools are best at producing completely correct content, but it also looks unmistakably weird. Adaptation means the AI rewrites the idea natively in Bahasa Malaysia based on the demand of the content type that you've specified. 

These are the six assets with their registers:

Asset 1: Landing Page

EN register: Professional tone, benefit-led
BM register: Formal ("anda" bahasa baku)

Asset 2: Email Copy

EN register: Warm professional
BM register: Semi-formal

Asset 3: FB/IG post

EN register: Conversational
BM register: Casual, light mix of rojak is fine

Asset 4: TikTok caption

EN register: Punchy
BM register: Colloquial ("korang", "jom")

Asset 5: AD Copy x 3

EN: Direct, adhere to policies
BM: Direct, formal

Asset 6: WhatsApp Broadcast 

EN: Short, personal
BM: Semi-formal, warm

You can change the assets to different content types, such as LinkedIn posts or X posts. Only the assets list can change, not the pipeline.

You can check out the Register Control Workflow, which serves as a documented register ladder.

Step 1: The Master Brief (10 minutes)

This is the most important part of your workflow. Be as detailed as possible.

MASTER CAMPAIGN BRIEF

1. AUDIENCE
   Who exactly:
   What they already know:
   What they're skeptical about:
   Language reality (EN-first / BM-first / mixed):

2. OFFER
   The product/service:
   The exact offer (price, discount, deadline):
   What's NOT included:

3. ANGLE
   The one idea this campaign leads with:
   Proof we can use (real numbers, real testimonials only):

4. BANNED CLAIMS
   Cannot say (e.g. "guaranteed results", "No. 1 in Malaysia"):
   Regulatory limits (KKM if health, BNM if finance,
   no "percuma" if conditions apply):

5. CTA
   Primary action + link:
   Real deadline (no fake scarcity):

Number 4 is the one often overlooked.

It helps keep you out of any trouble. When you're in niches like health, finance, or education, make sure it complies with regulations. 

I've also created a separate workflow on the banned-claims checker and included it with the locked drafts. 

The brief is prevention
The checker is detection

Step 2: The brand voice document (build once, reuse forever)

This one-page brand voice document is also used in the 2-Hour Brand Voice Document workflow.

== BRAND VOICE — [BRAND], v1, [DATE] ==

  • WHO WE SOUND LIKE:(one sentence)
  • WE ALWAYS:(3 rules)
  • WE NEVER: (banned words and habits, make sure to provide as much details possible)

LANGUAGE NOTES:

  • EN: (mechanics from step 2)
  • BM: (default pronoun per register; rojak policy; which terms stay in English)
  • ZH if you run the Chinese layer: point to that addendum]

  • 3 EXAMPLES OF US AT OUR BEST: (paste them in full)
  • 2 EXAMPLES OF WHAT WE'RE NOT: (one line each on why)

Please follow the voice document flow that I've linked. The pipeline for this workflow only requires that this document exist and be included in all prompts.

Step 3: English drafts (15 minutes)

One prompt that works for all six assets.

A single conversation in which your tone remains consistent across all assets.

Draft the following six assets in English.
The MASTER BRIEF and BRAND
VOICE document below are your only sources of truth.

  • Hard rules:
  • Every fact (price, number, deadline)  must appear in the brief.
  • The BANNED CLAIMS section is absolute. No paraphrasing around it.
  • Match the brand voice, including its "never" list.

1. LANDING PAGE: headline, subhead, 3 benefit blocks, CTA section.
2. EMAIL: 3 subject line options, ~150-word body, single CTA.
3. FB/IG POST: ~80 words, hook in the first line.
4. TIKTOK CAPTION: under 150 characters, up to 4 hashtags.
5. AD COPY ×3: primary text under 125 chars, headlines under 40 chars,
   three different angles from the brief.
6. WHATSAPP BROADCAST: under 100 words, personal tone, one link.

After all six, flag anything in the brief that was ambiguous or missing.

Step 4: Edit the English - before adapting (20-25 minutes)

Every fix that you make here propagates into Bahasa Malaysia for free. 

  • Make sure the facts match your brief
  • No banned claims slipped through (including paraphrasing)
  • The content should resonate with your brand voice with a clear CTA
  • Delete the weakest sentence in each of your assets or replace them with better ones

It's set now.

Lock your English edit.

Step 5: Bahasa Malaysia Adaptation (15 minutes)

These are the steps that separate your copy from translated ones.

Adapt the locked English assets into Bahasa Melayu.

This is ADAPTATION,not translation — rewrite each asset the way a native BM copywriterwould write it from scratch. Never translate idioms literally. 

Product names and technical terms stay in English. All banned-claim rules apply.

Register per asset:

1. LANDING PAGE: formal bahasa baku, "anda", zero slang.
2. EMAIL: semi-formal, warm, "anda".
3. FB/IG: natural social media BM, light rojak where a real Malaysian
   would use it.
4. TIKTOK: fully colloquial. "Korang", "jom". Reads like a creator
   wrote it, not a brand. Under 150 characters.
5. AD COPY: clean bahasa baku, policy-safe, same character limits.
6. WHATSAPP: semi-formal, reads like a person, not a blast.

Flag any line where you chose naturalness over literal accuracy.

If your Bahasa Malaysia comes out rigid without your brand flavor, the problem isn't "AI is bad at baku BM". The main issue is the incomplete register instructions. 

Step 6: The Native BM Pass (5 minutes)

The native speaker evaluates and goes through every asset.

We need a "Yes" answer for the following questions:

  • Does it sound like a Malaysian wrote it?
  • Is the register relevant for the platform?
  • Did any numbers, facts, or CTA drift from the English context?

Fix them inline and make sure they represent your brand voice. 

The Math

Manual or Hiring Agency

  • Turnaround: 2 days to 14 days depending on the complexity
  • Manual edits / human intervention: 12-16 hours
  • Cost per campaign: RM3,000+

This pipeline

  • Turnaround: Under 2 hours 
  • Manual edits / human intervention: 35-45 minutes
  • Cost:  Staff time + RM50-RM100/month for the AI tool of your choice

If you outsource your entire marketing campaigns to an agency, it's almost RM12,000+ a month. You can avoid this huge cost while using this workflow output repetitively and make necessary changes anytime to suit your brand voice.

Even if the person who built this resigns, the prompt in the AI tool's chat history won't go away anytime soon.

Where this workflow breaks

A lazy brief will result in confident rubbish in two languages. I know it's harsh, but that's the truth. The pipeline executes an angle; it doesn't automatically find one.

Editing after adaptation doubles the amount of correction work. Skipping the register instructions and blaming the model for BM that doesn't resonate with your brand voice will not do any justice. 

And if nobody in your team can do the native BM pass, you can hire a freelancer to run them through for RM50-RM100 per campaign.

You can build your master brief flawlessly by following the workflow step by step. 

PDFThe Master Brief Pipeline 6 Assets, 2 Languages, 2 Hours20 KB · PDFDownload                                                                               

Filed undermaster briefai for content assetsai for content development