Defining and applying a brand voice

0:00 Defining and applying a brand voice

Close your eyes and imagine a brand you love talking to you. You can probably hear it. The words it would use, the words it would never use, whether it would make a joke. That's brand voice. And when a brand has a clear voice, every touchpoint feels like the same trustworthy person. In this lecture you'll learn the difference between voice and tone, how to define a voice with the this-not-that method, how to turn a voice guide into a working system that people and AI tools use every day, and how voice survives translation.

0:42 Why voice matters for conversion

Why does voice matter for conversion, not just branding? Because consistency builds trust, and trust converts. If your ads sound playful, your website sounds corporate and your emails sound robotic, readers subconsciously wonder whether it's the same company. A clear voice also speeds up work. Writers, freelancers and AI tools make fewer wrong choices when they know the rules. And with AI drafting more copy than ever, voice is what stops your brand from sounding exactly like everyone else using the same model.

1:18 Voice vs tone

Voice and tone are different. Voice is your personality, and it stays the same. Tone is how that personality adapts to the situation. Think of a good doctor. Their personality is the same whether they're sharing good news or bad news. But their tone changes. Warm and upbeat for the first, calm and careful for the second. Your brand might be warm, expert and direct. In a launch email, that sounds confident and excited. In an apology after a delivery failure, it sounds calm, responsible and straightforward.

1:56 The 'this, not that' method

Here's how to define a voice. Use the this, not that method. Pick three or four traits and define each with its boundary. Warm, not gushing. Expert, not academic. Direct, not blunt. Playful, not silly. The not half is what makes it usable, because it tells writers where the line is. Then add examples: a sentence that nails each trait, and one that crosses the line. Base the traits on research: how customers describe you, what they value, and how your best-performing copy already sounds.

2:33 Simple example: freelancer fintech app

A simple example. A fintech app for freelancers in Pakistan and the Gulf defines its voice as clear, not technical; reassuring, not patronising; and practical, not preachy. Here's how that changes a push notification. Before: transaction number four eight two nine has been successfully processed. After: you've been paid. Fifty thousand rupees from Studio Nine just landed in your account. And an error message. Before: an error has occurred. After: we couldn't reach your bank just now. Your money is safe. Try again in a few minutes. Same facts, but now it sounds like the brand.

3:15 Realistic example: UK and Gulf skincare (illustrative)

Now a realistic scenario with illustrative details. A skincare brand selling across the UK and the Gulf has five freelancers and two AI tools producing copy. Customers say the brand feels inconsistent: jokey on TikTok, clinical on the website, pushy in emails. The team builds a voice system, not just a guide. Four traits with this, not that examples. A word list with product names, banned superlatives, British spelling and how to write prices in pounds and dirhams. A tone map for launches, apologies and errors. Fifteen gold-standard examples. And a reusable instruction block loaded into their AI tools' project settings. Within a few weeks, the copy review queue shrinks, because first drafts arrive much closer to the voice.

4:07 Watch me do it: AI voice instruction block

Watch me build and test the AI instruction block. I start with who we write for and the markets. Then voice traits in this, not that form. Then the word list: customers, not users; plan, not package; British spelling; prices written as AED forty-nine a month. Then the never list: unproven superlatives, clichés like unlock and seamless, emojis in email, health outcome claims. Then the tone map, and a line telling the model to use only facts from the brief and write TBC if something's missing. Now I test it with three briefs: an ad, an apology email and an error message. I compare each output with my gold-standard examples, fix the instructions where it drifts, and save it as version one point one with today's date.

5:02 Voice in everyday situations

Let's look at voice in everyday situations, because that's where it's won or lost. A delivery is late. Off-voice: we apologise for any inconvenience caused. On-voice for a warm, direct brand: your order is running a day late, and that's on us. It'll arrive Thursday, and we've refunded your delivery fee. A customer asks a simple question on Instagram. Off-voice: please refer to our FAQ. On-voice: yes, it's machine washable, just keep it under forty degrees. A price goes up. Off-voice: due to market conditions, prices have been adjusted. On-voice: from next month our plan costs a little more, here's exactly why and what's staying the same. Voice shows up most clearly when things are awkward.

5:52 Where the voice system lives

Where does the system live? The voice principles and word list sit at the top of your message bank. Gold-standard examples live in a shared folder. The instruction block is loaded into your AI tools. Most major assistants let you save reusable instructions and reference files, for example Claude Projects, ChatGPT Projects or custom GPTs, and Gemini Gems. Writing tools like Jasper, Writer and Grammarly offer brand voice or style guide features. And the review checklist names who approves which content type. The point isn't the tool. It's that the same rules travel everywhere copy is made.

6:34 Voice across languages

Voice must survive translation. If you write in English, Arabic and Urdu, define traits as behaviours, like we explain before we sell, rather than English wordplay that can't be translated. Have native writers create examples in each language, instead of translating English lines word for word. Formality levels differ. Humour lands differently. And religious and cultural references need local judgement, especially around Ramadan and Eid. A consistent voice across languages feels like the same person speaking fluently, not a phrasebook.

7:09 Common mistakes

Common mistakes. Voice traits that everyone claims, like friendly and professional, with no boundaries. A guide with no examples. Confusing voice with tone, so apology emails sound as jokey as launch posts. A guide that lives in a PDF nobody opens. AI tools with no voice instructions, so every draft sounds generic. And translating voice word for word. Each makes your brand sound like several different companies.

7:38 Recap

Recap. Voice is your stable personality. Tone adapts to the moment. Define voice with three or four this-not-that traits, backed by examples. Turn it into a system: word list, tone map, gold-standard examples, AI instructions and a review checklist, and make sure it travels to every tool and language. Try this now: write your this-not-that traits, then build the instruction block from the lesson text and test it with three briefs: an ad, an apology and an error message. Compare the outputs with your best existing copy, and refine until the drafts sound like you.