How to Write Brand Voice Guidelines | Presspage

Brand voice is a funny thing. It can be hard to pinpoint when it’s off, and for it to work across an organisation, a lot of people need to be fluent in it.

Sometimes, a draft will land in your inbox and, on the surface, everything looks fine. The grammar is clean, the facts are all there, and there’s nothing obviously ‘wrong’. But still, it just ‘doesn’t sound like us’.

Brand voice guidelines are a practical set of rules that define how your organisation should sound in writing. Good guidelines give everyone involved in creating and reviewing content a shared reference point for tone, vocabulary, style, and boundaries.

With AI entering the chat in a big way, getting that reference point right has become even more important. The 2026 State of AI in PR report found that 74% of PR professionals use AI to write and create content, while 86% use it to edit and refine their work.

So how do you write brand voice guidelines that people can actually use, while giving AI enough direction to stay recognisably on-brand?

Let’s get into it.

Brand voice vs. tone: what is the difference?

Brand voice and tone are closely linked, but they do different jobs.

Your brand voice is the recognisable personality behind your communication. It should stay relatively stable whether you are writing a press release, answering a journalist, or publishing a LinkedIn post.

Your tone changes with the moment. A product launch can sound upbeat, whereas a crisis response will likely need more empathy and precision. The person speaking hasn’t changed, but the situation has.

 

Poor brand voice guidelines often condense the two. A brand may say it wants to sound “friendly,” then apply the same breezy tone to a launch announcement, redundancy notice, customer complaint, and security incident. That doesn’t work.

What should you include in your brand voice guidelines?

A one-page list of adjectives won’t carry a global comms team very far. Your guidelines need enough detail for someone (potentially someone with little knowledge of your brand) to make a decision while they are drafting.

At minimum, cover these areas:

1. Brand purpose and positioning. Explain the communication idea behind the voice. What does the organisation stand for? What should audiences understand about you after reading your content?
2. Voice and tone. Define what stays fixed and what can shift situationally.
3. Three to five personality traits. Give each trait a boundary such as “confident, not arrogant” or “warm, not chatty.”
4. Audience profiles. Show how the tone may change for customers, journalists, employees, and other groups.
5. Preferred vocabulary. Keep a record of approved terminology, product names, phrases, and recurring language.
6. Words and phrases to avoid. Include jargon, banned terminology, and wording that clashes with the brand.
7. Do’s and don’ts. Give side-by-side rewrites so people can see the rule in practice.
8. Grammar and style rules. Cover spelling, punctuation, capitalisation, dates, numbers, headline style, abbreviations, and formatting.
9. Channel and format variations. Explain what changes across press releases, social posts, newsroom stories, and other channels.
10. Tone by situation. Define how the voice behaves across different scenarios such as everyday communication, good news, difficult news, complaints, etc.
11. Crisis communication rules. Set expectations for holding statements, confirmed facts, spokesperson language, authorship, and approval.
12. AI guardrails. State clearly where AI can assist, what information should never be entered, what checks are mandatory, and which content always needs human review.
13. On-brand and off-brand samples. Give people finished examples to compare against, ideally using real content types from your organisation.

 

Brand voice guidelines examples

Airbus: Turn the rule into a rewrite

Airbus anchors its tone of voice in its company purpose and three broader ‘imperatives’: unite and collaborate, care and respect, and influence and pioneer. From there, it gets very practical.

The strongest part is the use of before-and-after examples.

For active voice, Airbus gives writers “We guarantee production of the twin-engine jet fighter at least until 2030” instead of the passive “Production of the twin-engine jet fighter is assured at least until 2030.” It also tells writers to speak directly to stakeholders and to choose familiar words over inflated language.

Airbus also offers very clear guidelines on inclusive language, instructing users to use gender-neutral language such as ‘Hello Everyone!’ instead of ‘Hello Guys’ and to be mindful of cultural differences by wishing ‘Happy Holidays’ instead of ‘Merry Christmas’.

One line sums up the Airbus approach neatly: “We don’t make things bigger than they are. We say it how it is.”

That is a much stronger instruction than “sound clear.” It gives the writer a behaviour to follow and a habit to avoid.

What to borrow: For every major voice rule, add a real sentence that follows it and one that breaks it. If a writer cannot see the difference, the rule needs more work.

 

TC Energy: define traits with boundaries

TC Energy gives its voice three core traits: “confident, approachable and optimistic.”

The practical part comes immediately after the labels, with each trait getting a short explanation of what it means in practice.

“Confident,” for example, means speaking with clarity and conviction without tipping into arrogance. “Approachable” means keeping technical communication conversational and avoiding insider jargon. “Optimistic” is tied to action rather than empty positivity.

These examples are the key, as many brand voice documents stop at adjectives. “Friendly. Bold. Human.” That’s great. But what happens when two writers interpret “bold” in completely different ways? A trait only becomes usable when it has defined edges.

What to borrow: Give every trait a “this, not that” boundary and a short explanation of how it changes the writing. “Confident, not arrogant” tells a writer far more than “confident” on its own.

 

NHS England: define how tone changes when the stakes change

NHS England’s tone of voice rules are a strong example of situational tone. Its general writing principles ask for communication that is clear, concise, straightforward, honest, open, professional, respectful, and accessible.

The rules change when the subject becomes more sensitive.

For difficult topics, NHS England says: “your tone should communicate genuine understanding and respect.” The aim is for the reader to feel informed and empowered.

That distinction is crucial for PR teams. The brand voice still has to be recognisable, but the emotional setting changes. Humour that works in a campaign post may be wildly inappropriate in a public safety update. High-energy language that suits a launch can sound evasive when people need facts.

What to borrow: Include a tone matrix for high-stakes situations. Show what stays fixed, what gets dialled down, and what becomes more prominent.

How to write brand voice guidelines

You don’t need everything all at once, so start small with the decisions writers make most often, then build from there.

Step 1: Pick your personality and tone traits

Choose three to five traits that describe how the brand should come across. Fewer tends to work better because people can actually remember them.

Then add boundaries.

Instead of:

Clear

Try:

Clear, not clinical
We explain complex subjects in everyday language. We keep the detail people need, skip the internal jargon, and avoid sounding like we swallowed the policy PDF.

Instead of:

Confident

Try:

Confident, not self-important
We state what we know directly, and we back claims with evidence. We choose more humble language and avoid inflated superlatives.

Instead of:

Warm

Try:

Warm, not overly familiar
We write like a person speaking to another person. We can use contractions and natural phrasing, but we do not force jokes, slang, or fake emotionality.

Step 2: Build your vocabulary and do’s and don’ts

Create a vocabulary section with three buckets: preferred language, language to avoid, and terms that need context or approval.

For example:

 

Then add before-and-after examples from the formats your team writes most often.

Step 3: Plan for edge cases, including crisis tone

Everyday comms is easy - the real test of your brand voice guidelines comes when the situation changes.

Identify the moments when tone needs to shift (a crisis, outage, product recall, public criticism, etc.) In those moments, the voice can stay recognisable while certain traits move forward or back. A playful brand may become restrained; an energetic brand may become calmer.

This is where a crisis tone guide helps. Define:

• which traits remain fixed
• which traits should be reduced or paused
• which words or claims require approval
• how uncertainty should be expressed
• who can draft a holding statement
• who approves it
• which spokespersons can speak publicly
• when the next update should be promised

Our own Crisis Comms Guidelines recommend preparing holding statements in advance and keeping early communication focused on confirmed information, care for people affected, what the organisation is doing, and when further updates will follow.

Our top tip? Set the core writing and approval group before you need it. A crisis draft that passes through twelve writers can quickly lose clarity and accountability, so reducing the number of cooks in the kitchen up front can be a big help.

Step 4: Put the guidelines where people need them

If no one has visibility on your brand voice guidelines, people will fill in the gaps themselves.

AI adds another layer to that. A March 2026 survey of 858 marketing and PR professionals found that only 21.4% of organisations had formal, documented AI policies. Nearly a third, 29.2%, named loss of brand voice as a top concern, behind data privacy and accuracy.

And the concern is valid. You can give an AI tool a PDF full of brand rules, but that does not mean those rules will be followed every time someone starts a new prompt, rewrites a paragraph, or adapts a story for another market.

The same problem applies to people. If the latest terminology, messaging, and tone rules are not front-and-centre during the writing process, someone will eventually work from the wrong version.

So the job does not end when the guidelines are written. Put them where people are actually creating and reviewing content. Keep approved language in one place, make the checks clear, and make sure teams, agencies, and AI tools are all working from the same set of rules.

That moves brand voice from a branding exercise into part of the comms workflow.

What are brand voice guidelines?

Brand voice guidelines define how an organisation should sound in written communication. They usually cover personality traits, tone, vocabulary, words to avoid, editorial style, channel differences, situational tone, and examples of on-brand and off-brand writing. Strong guidelines help different writers make similar language choices without needing the same person to rewrite every draft.

What should brand voice guidelines include?

At minimum, include your brand purpose, three to five voice traits, a definition of voice versus tone, preferred vocabulary, language to avoid, editorial rules, do’s and don’ts, channel examples, tone for difficult situations, AI guardrails, and real sample copy. If your organisation operates across markets or languages, include rules for localisation and approved terminology too.

What is the difference between brand voice and tone?

Brand voice is the personality behind your communication. Tone is how that personality adapts to a specific context. Your voice might be clear, candid, and warm. Your tone could become celebratory for a launch, reassuring during an outage, or restrained during a crisis.

How many brand voice traits should you have?

Three to five is a practical range for most organisations. Each trait should have a clear boundary and an explanation of what it changes in the writing. Five specific, well-defined traits are far more workable than fifteen adjectives nobody can remember.

How do you keep AI-generated content on brand?

Start by giving AI the same information you give a human writer: approved messaging, voice traits, preferred terminology, banned terms, style rules, and examples. Then define which content requires human review and what the reviewer needs to check. For teams producing large volumes of content, brand governance software can apply those rules during drafting and flag language that falls outside the approved standard before publication.

 

The takeaway

If your team is having trouble recognising off-brand content, your voice rules are probably too vague or too far away from the work.

Write guidelines people can act on. Define a small set of traits, define the boundaries, map the vocabulary, show before-and-after examples, and plan for the moments when the tone needs to change. Then put those rules into the workflow used by your people, agencies, and AI tools.

You’ll know the guidelines are doing their job when drafts start arriving already on-brand.

Ready to move brand voice from a PDF into the way your team actually works? Explore Presspage Brand Governance >>>