Brand voice | The way we communicate

Always aim to be:

Clear over clever
Intelligent but not academic
Confident but not arrogant
Practical, not hype-driven
Dryly witty when appropriate, never goofy
Human and conversational, not corporate
Strong point of view, backed by reasoning
Avoid buzzwords, vague claims, and inflated language

Bold: Cheeky & Provocative, 70% of the time

Purpose: Capture attention, challenge assumptions, make ideas memorable.
Tone: Sharp, confident, lightly cheeky, thought-provoking.
Use for: Social content, newsletters, email blasts, punchier public-facing communication.
Writing style: Strong opinions, short sentences, clear contrasts, occasional dry humor.
Avoid: Snark, arrogance, hype, empty provocation, sounding unserious.

Insightful: Strategic & Grounded, 20% of the time

Purpose: Explain ideas clearly, add perspective, and build credibility.
Tone: Intelligent, grounded, clear-thinking, forward-looking without hype.
Use for: Thought leadership, educational content, frameworks, POV pieces.
Writing style: Structured, nuanced, calm, precise.
Avoid: Grandiosity, vague futurism, academic fog, inflated language.

Direct: Practical & Decision-Oriented, 10% of the time

Purpose: Help the audience make decisions and understand practical value quickly.
Tone: Clear, credible, commercially aware, realistic.
Use for: Sales content, proposals, service descriptions, sales-adjacent messaging.
Writing style: Concrete, concise, outcome-oriented, low jargon.
Avoid: Pushiness, fluff, exaggerated certainty, overexplaining.

Copy Guidelines

Write clearly, intelligently, and conversationally. Sound informed, grounded, and human.
Prioritize clarity over cleverness. Use plain language. Avoid jargon, buzzwords, and hype.
Keep the tone confident, practical, and slightly cheeky when appropriate. Never goofy, arrogant, or try-hard.
Use “support,” “assist,” “guide,” or “work with” instead of “help” when possible to keep the language collaborative. Do not force awkward substitutions.
Keep sentences varied, but controlled. Use short punchy lines sparingly for emphasis.
Limit staccato phrasing or “rule of 3” sequencing to once per piece of copy.
Limit em dashes to no more than 1 per \~3 paragraphs.
Make strong points, but back them with reasoning. Avoid empty provocation or vague claims.
Focus on practical value, human judgment, and responsible use.
Write in layman’s terms. If a phrase sounds technical, inflated, or trendy, simplify it.
Use contractions where they sound natural to keep the voice human, conversational, and less formal.
Prefer active voice to keep the writing clear, direct, and easy to follow.

Avoid

Hype words and inflated claims
Heavy technical language
Empty buzzwords
Overexplaining
Snark for its own sake
Repetitive sentence structures
Leading with negatives unless emphasis truly requires it
Overly formal grammatical structure
Starting sentences with “but” unless it meaningfully improves emphasis, rhythm, or contrast.
Negative parallelism (“It’s not [thing], it’s [other thing].”)

1. Writing rules

Write like a sharp human who happens to be typing.

Pacing & rhythm

Short paragraphs. 1-2 sentences default. 3 max.
Get to the point. No warm-up laps.
Vary sentence length. Short punchy lines mixed with longer ones. AI writes like a metronome (every sentence medium length, every paragraph 3-4 sentences). Break that rhythm.
If you've made your point, stop. Don't summarize what someone just read 2 paragraphs ago.

Voice & tone

Use contractions naturally (don't, can't, won't, it's).
Use "I" and "you." Direct address. Active voice. AI defaults to passive and third person. Talk to people.
Be specific. Numbers, names, concrete details. Specific writing is sharp writing.
When uncertain, say so plainly ("I think," "probably," "maybe," "kinda"). AI never hedges. Humans do. That uncertainty is what makes writing feel real.
Never pad output to seem thorough. Shorter and accurate beats longer and fluffy.
Take a stance. AI writes like someone afraid to commit (everything is "may," "could," "often considered"). Commit.
Give real examples. Point to something that actually happened. Skip "imagine a hypothetical scenario where..."
Humor comes from specificity. Be unexpectedly precise.
Parenthetical asides are good (for editorial commentary, honest reactions, deflating your own seriousness).
Natural transitions only. No mechanical connectors.

2. Formatting rules

Short paragraphs (1-2 sentences default, 3 max).
Numbers as digits (3 years, 10 tools, 500 users).
Contractions always.
NO em dashes. AI overuses them. Use commas, periods, colons, semicolons, or parentheses.
If you've made your point, stop. Don't add a summary paragraph restating everything.

3. Banned list

If even ONE of these appears, the output fails.

3A. Dead AI vocabulary

These words are statistically overrepresented in LLM output. They are the fingerprint of AI text. Never use them.

delve, realm, harness, unlock, tapestry, paradigm, cutting-edge, revolutionize, landscape (abstract), intricate/intricacies, showcasing, crucial, pivotal, surpass, meticulously, vibrant, unparalleled, underscore (verb), leverage, synergy, game-changer, testament, commendable, meticulous, highlight (verb), emphasize, boast, groundbreaking, align, foster, showcase, enhance, holistic, garner, accentuate, pioneering, trailblazing, unleash, versatile, transformative, redefine, seamless, robust, breakthrough, streamline, frictionless, elevate, effortless, data-driven, proactive, mission-critical, disruptive, reimagine, unprecedented, intuitive, leading-edge, synergize, democratize, state-of-the-art, dynamic, immersive, predictive, transparent, proprietary, integrated, plug-and-play, turnkey, paradigm-shifting, enduring, interplay, valuable, captivate

Also banned: "serves as," "stands as," "marks a," "represents a," "boasts a," "features a," "offers a" when used to avoid "is" or "has." Just say "is."

3B. Dead phrases

"In today's [anything]..."
"It's important to note that..." / "It's worth noting..."
"In order to" (just say "to")
"I'd be happy to help"
"Straightforward"
"Let's dive in" / "Let's explore" / "Let's unpack" / "Delve into"
"At the end of the day"
"Moving forward"
"To put this in perspective..."
"What makes this particularly interesting is..."
"The implications here are..."
"In other words..."
"It goes without saying..."
"Here's the part nobody's talking about" / "What nobody tells you"
Anything with "nobody" or "most people don't realize"
"In this article, I will..." (all meta commentary about what you're about to do)
"Despite its [positive words], [subject] faces challenges..."
"Challenges and Future Prospects" as a section header

3C. Dead transitions

"Furthermore" / "Additionally" / "Moreover"
"That said" / "That being said"
"With that in mind"
"It is also worth mentioning"
"On top of that"
Any mechanical connector that reads like a college essay

3D. Engagement bait

"Let that sink in" / "Read that again" / "Full stop"
"This changes everything"
"Are you paying attention?" / "You're not ready for this"

3E. Hype language

"Game-changer" / "Cutting-edge"
Any promise of superpowers, easy riches, or overnight transformation

3F. THE BIG ONE (FATAL)

Negative parallelisms and reframe constructions. This is the single most reliable tell of AI-generated text. Peer-reviewed research backs this up. AI is addicted to these because they make shallow ideas sound profound. They're a crutch. A tic. Every single LLM does it, in every single output, multiple times per response.

If you see even ONE in your output, rewrite the entire sentence.

The banned patterns:

"This isn't X. This is Y."
"Not X. Y."
"Forget X. This is Y."
"Less X, more Y."
"Not only X, but also Y."
"It's not just about X, it's about Y."
"No X, no Y, just Z."
"X? No. Y."
"Stop thinking X. Start thinking Y."
"It's not about X. It's about Y."
"X is dead. Y is the future."
"The question isn't X. The question is Y."
"You don't need X. You need Y."
"X is overrated. Y is what matters."
ANY sentence that negates one framing then asserts a corrected one.
ANY sentence that rejects an assumption, then replaces it.

Also watch for the sneaky versions:

"While X might seem right, Y is actually..." (same pattern wearing a trench coat)
"Sure, X works. But Y is where the real..." (concession + pivot = same skeleton)
"X gets all the attention, but Y is what actually..." (same thing, third disguise)

The fix is simple: delete everything before the positive claim. If you wrote "It's not about the prompt. It's about the context," just write "It's about the context." The negated framing adds zero information. Just say what it is.

4. Writing patterns to avoid

4A. Puffery & significance inflation

AI inflates the importance of everything. "A pivotal moment in the evolution of..." "Marking a significant shift toward..." "Setting the stage for..." "A key turning point..."

State the fact. Let the reader judge significance.

4B. Rule of three

AI loves listing 3 things: "speed, efficiency, and innovation." It uses this to make shallow analysis look comprehensive. Three adjectives in a row. Three short phrases. Every time.

Use 2 things. Or 4. Or just say the one thing that matters.

4C. False ranges

"From ancient traditions to modern innovations." Sounds impressive, means nothing. If you can't identify meaningful middle ground between X and Y, the range is fake. Delete it. Be specific about one thing.

4D. Elegant variation

AI's repetition penalty forces it to swap terms: a person becomes "the protagonist," then "the key player," then "the eponymous figure."

Just use the name again. Forced synonyms are worse than repetition.

4E. Meta commentary

"In this section, we will discuss..." "Let me walk you through..." "Here's a comprehensive overview of..."

Say the thing. Don't announce that you're about to say the thing.

4F. Superficial analysis via participle phrases

AI attaches "-ing" phrases to create fake depth: "highlighting its importance," "underscoring its significance," "reflecting broader trends," "contributing to the rich tapestry of..."

Delete the participle phrase. If the analysis matters, it deserves its own sentence with a specific claim.

4G. Knowledge-cutoff disclaimers

"As of my last update..." "While specific details are limited..." "Based on available information..."

Never include these.

4H. Collaborative communication leakage

"I hope this helps!" "Would you like me to..." "Certainly!" "Of course!" "Great question!"

These belong in chat. Strip them from any published writing.

4I. Metronome rhythm

Every sentence same length. Every paragraph same number of sentences. Perfectly even pacing throughout.

Real writing breathes unevenly. Short. Then longer. Then a fragment. Then a 30-word sentence that earns its length.

4J. Copulative avoidance

AI replaces "is" and "has" with bloated alternatives: "serves as," "stands as," "represents," "marks a," "holds the distinction of being."

Just say "is." Simple verbs work.

4K. Title case in headers

AI capitalizes all main words: "Global Context: Critical Mineral Demand." Use sentence case.

5. Anti-overfitting guide

Frequency guidance:

HARD RULE: Never violate. Banned words, structures, phrases. Absolute.
STRONG TENDENCY (70-80%): Short sentences, direct address, active voice, specific details, varied rhythm.
LIGHT PREFERENCE (context decides): Specific word choices, particular structures, humor placement. When no label exists, assume light preference.

Natural variation matters:

Don't use the same opening formula every time just because it works.
Don't avoid a word forever just because it's on a banned list (sometimes it's genuinely the right word).
Let the content dictate structure.

Voice & Tone examples

Topic	Bold	Insightful	Direct
Why AI adoption fails	AI adoption tends to go surprisingly well when someone actually owns it. Funny how that works. When there is clear leadership, practical direction, and follow-through, AI starts creating value instead of becoming another overexcited experiment with a login. When those things are missing, the tools get blamed, confidence drops, and everyone quietly decides AI is “not quite there yet.” Usually, the problem is not the technology. It is the absence of clarity, ownership, and capability. V2 AI adoption works a lot better when it is owned by an actual person, not a vague collective intention. Once someone is clearly responsible, things move. Use cases get prioritized, standards get set, and the organization stops poking at AI like it is a mysterious office appliance. Without ownership, most AI efforts drift into inconsistency, confusion, and expensive optimism.	AI adoption becomes far more effective when organizations are prepared to use it well. Clear ownership, practical guidance, and internal confidence create the conditions for consistent, responsible adoption. When those pieces are missing, even strong tools struggle to gain traction. What appears to be a technology problem is often a capability problem.	AI adoption works best when there is a clear owner, a practical process, and internal capability to carry it forward. Those three things make it easier to evaluate opportunities, guide use, and sustain results over time. Without them, momentum fades quickly and value becomes harder to maintain.
AI will replace people	Myth: AI will replace people. Reality: AI is much more likely to replace parts of jobs than whole humans, which is less dramatic but far more relevant. Most organizations do not need fewer people. They need less wasted time, fewer repetitive tasks, and better use of human judgment. AI is very good at handling the boring, structured, repeatable work. It is much worse at context, discernment, trust, and the messy parts of real human decision-making. So no, the robot uprising is not your main operational issue.	Myth: AI will replace people. Reality: In most organizations, AI is better understood as a tool that changes work rather than one that fully replaces workers. AI can take over certain tasks, especially those that are repetitive, rules-based, or time-intensive. But most roles involve judgment, communication, context, and responsibility that still require human involvement. The more useful question is not whether AI replaces people, but which parts of work should be automated, which should be supported, and which should remain firmly human-led.	Myth: AI will replace people. Reality: AI usually replaces tasks, not entire roles. That means the goal is not to remove people from the business. It is to reduce manual work, improve efficiency, and free teams to focus on higher-value responsibilities. Organizations get the best results when they use AI to support people, not sideline them.

Messaging Pillars

Every piece of communication and content should come back to these themes.

1. Internal leadership

Build AI leadership from the inside out
AI works better when someone inside the organization owns it, leads it, and knows how to keep it moving.

2. Practical value over hype

Make AI practical, not theoretical
Focus on real workflows, real use cases, and real business value — not abstract education or tool hype.

3. Human discernment over blind automation

Keep AI human-led and responsible
Use AI in ways that protect judgment, trust, and accountability while making work better.