Skip to main content

Brand voice & personas

Why this matters for your business​

The single biggest reason AI content sounds bad is the AI doesn't know your brand. Generic "professional" copy is the default; it takes 30+ minutes of editing per piece to make it sound like you instead of ChatGPT. That editing time eats most of the AI's time savings.

Brand voice + personas fix this at the root. You capture the brand's voice once — through samples, do/don't lists, tone sliders, and target audiences. Every generator on the platform (blog, ad, email, video, image) uses that captured voice as input. Output sounds like your brand on first generation, not after editing.

This is the highest-leverage 30 minutes you'll spend on the platform. Skip it and AI content disappoints; do it and AI content compounds.

What this typically unlocks​

OutcomeResult
Edit time per generated piece−70% typical
First-pass usability of AI output70-85% vs. 20% generic
Brand consistency across surfacesmeasurable +
Time to set up30-45 minutes one-time

What you actually capture​

Brand voice​

Three components:

  1. Tone sliders — formal ↔ casual; serious ↔ playful; reserved ↔ enthusiastic; expert ↔ approachable. Set 5-10 sliders to position your brand.
  2. Sample library — paste 3-5 examples of "perfect" brand copy (existing emails, product descriptions, social posts). The system learns from samples.
  3. Do / Don't lists — specific words / phrases to use ("crafted", "founder-led", "small-batch") and avoid ("revolutionary", "game-changing", "synergy").

Personas​

Define your top 2-4 customer personas. Each:

  • Name + brief description
  • Demographics (age, location, life stage)
  • Motivations (what they buy your product for)
  • Objections (what stops them from buying)
  • Voice match (how they speak — affects what tone resonates)

Personas let generators target specific audiences ("write this ad for Persona 1: time-strapped young professional").

Real merchant scenarios​

Scenario A — Brand voice captured in 30 minutes​

Setup. Beauty brand. Founder spent 30 min:

  • Set 8 tone sliders
  • Pasted 5 of her favorite emails
  • Listed 12 brand do's + 8 don'ts
  • Defined 3 personas (skincare-newbie, routine-builder, ingredient-conscious)

Effect on first generated email:

Before brand voice setupAfter
Edit time25 min8 min
Tone match4/108/10
Founder confidence shipping as-israre~70% of pieces

Scenario B — Mid-market brand maintains voice across team​

Setup. Marketing team of 4. Pre-platform: each writer had slightly different tone; brand felt inconsistent.

With brand voice captured centrally:

  • All 4 writers use AI generation as starting point
  • Output has same voice regardless of who's writing
  • Brand consistency improved measurably (customer survey: "the brand feels coherent" went 6.2 → 7.8 / 10)

Scenario C — Persona-targeted ads outperform generic​

Setup. Brand has Persona 1 (price-conscious) and Persona 2 (premium-buyer).

A/B test: Generic ad copy vs. Persona-1-targeted vs. Persona-2-targeted, served to corresponding audiences.

Result:

  • Generic: 1.4% CTR
  • Persona-targeted: 2.6% CTR (+86%)

The persona definition is what made the targeted copy resonate.

Best practices​

✅ Spend the 30 minutes setting up brand voice first. Before generating anything else.

✅ Update brand voice quarterly. Brand evolves; samples should reflect current state.

✅ Define 2-4 personas, not 10. More dilutes targeting.

✅ Use do/don't lists liberally. Specific words you love or hate are the strongest signals.

❌ Don't skip the sample library. Tone sliders alone are weak signal; samples are strong.

❌ Don't rely on generic AI without brand voice. Output will disappoint and you'll blame the wrong thing.

Plan tiers​

CapabilityFreeStarterProAgencyEnterprise
Brand voice capture✓✓✓✓✓
Persona definitions (up to 4)✓✓✓✓✓
Sample library✓✓✓✓✓
Multi-brand voice (per shop)———✓✓
Voice analytics ("which voice variant won?")——✓✓✓
Cross-brand voice library———✓✓

See also​