| name | brand-visual-identity |
| description | Use when defining or refreshing a distinctive visual identity across logo usage, colour, type, spacing, imagery, and applications. Unlike logo-and-wordmark-design, this owns the whole identity system; detailed palette and type construction route to specialist skills. |
| metadata | {"portable":true,"category":"02-color-brand-and-visual-identity","compatible_with":["claude-code","codex"]} |
Brand Visual Identity
Acknowledgement: Shared by Peter Bamuhigire, techguypeter.com.
Use When
- Defining a new brand's visual identity from scratch: logo, colour, type, imagery, and the
rules that bind them into one system.
- Refreshing or auditing an existing identity that has drifted into generic "startup sameness"
— stock gradients, an interchangeable geometric-sans wordmark, swappable hero illustration.
- Producing a mini style-guide / brand sheet that downstream artifacts (site, app, deck, DOCX
suite) must obey for consistency.
- Deciding logo placement, lockups, clear-space, and minimum sizes for a specific surface.
Do Not Use When
- The task is only choosing typefaces — use
font-selection-and-pairing and return here to
tie the choice into the identity.
- The task is only constructing or correcting a colour palette / contrast — use
color-system-and-palette and return here for the system-level tie-in.
- The task is auditing imagery/UI for generative-AI artefacts (warped logos, uncanny photos) —
run
skills/00-cross-cutting-ops-qa-a11y/visual-product-slop-audit first, then design the
replacement here.
- A binding client brand guideline already exists — follow it, record it, and apply only the
consistency gate.
Required Inputs
| Input | Supplied by | Required? | Why |
|---|
| Brand strategy, audience, promise, and differentiation | Brand owner/strategy | yes | Anchors every choice |
| Existing assets, equity, and constraints | Client or source system | conditional | Controls preservation |
| Required channels and applications | Delivery brief | yes | Proves the system works |
- Who the brand is for and against: audience, the category it competes in, and the two or
three competitors it must NOT be confused with. Distinctiveness is relative; you cannot be
ownable without naming the field you stand apart from (Neumeier, The Brand Gap — "a brand
is not what you say it is, it's what they say it is"; differentiation is the first job).
- One-word brand essence and 3–5 adjectives (e.g. "austere, precise, warm") that every visual
decision is tested against.
- The surface set the identity must survive: favicon → billboard, light/dark, print/screen,
embroidered/single-colour.
- Any existing assets, and whether a logo already exists or must be commissioned/designed.
Workflow
-
Fix the strategy before any mark. Write the one-word essence and the adjective set, and
the "not-to-be-confused-with" list. Every later choice is justified against these — this is
the human-authority anchor that stops the identity converging on the category mean
(doctrine Mission: the moat is looking human-made; doctrine/design-doctrine.md §0).
-
Logo / wordmark — make it ownable, not generated.
- Prefer a wordmark or monogram built from real, correct letterforms over an abstract
blob. If a custom mark is drawn, it must be drawn deliberately — Paul Rand's discipline:
the mark earns meaning through use, so it must be simple, distinctive, and reproducible
at any size, in one colour, and from memory.
- Never ship an AI-generated mark. Generative marks carry the slop tells — warped or
misspelled letterforms, fused glyphs, melted detail, near-symmetry that reads as "off"
(
doctrine/references/ai-slop-taxonomy.md §Visual tells; category 2, "misspelled logos /
uncanny promo art" is the highest public-backlash risk for a brand).
- Specify the system, not one image: primary lockup, secondary/stacked lockup, monogram,
clear-space (defined in units of the mark itself), and minimum legible size verified at
the favicon end. Vignelli's discipline applies: a small, governed set of elements used
consistently, not a sprawl of one-off variants.
-
Tie in colour. Do the palette work in color-system-and-palette, then bind it: assign
roles (brand / ink / surface / accent / state), and choose one signature colour the
brand can own — not a purple→blue gradient, which is the single most recognisable identity
slop signal (doctrine §2). State why this colour fits the essence.
-
Tie in type. Do the pairing in font-selection-and-pairing, then bind it: the display
face is part of the identity's voice, not just headings. Record where the wordmark's
letterforms and the running type agree or deliberately contrast (a wordmark may be custom
while body uses the paired body face — say so).
-
Define the supporting visual language — these are what separate an authored identity
from a template:
- Spacing & geometry: a consistent spacing rhythm and corner/edge language (sharp vs
soft) applied everywhere. Cross-ref
skills/03-layout-grid-and-composition/layout-grid-and-spacing for the grid.
- Photography style: state the rules — framing, colour grade, subject treatment, whether
people are real. Reject uncanny / waxy / impossible-physics AI imagery as brand
content (
ai-slop-taxonomy.md §1–2).
- Illustration / iconography style: one coherent system (line weight, corner radius,
perspective) — not a marketplace of mismatched stock icons.
- Motion (if any): one or two signature behaviours, not decoration.
-
Write the voice-of-the-visuals. One short paragraph + a do/don't pair that tells any
future contributor how the brand behaves visually — the gap Neumeier names between strategy
and execution. This is the rule a templating tool can never infer.
-
Produce the mini style-guide deliverable (see Outputs) and run the Consistency Gate.
Structure the system against references/identity-system-spec.md; for a complete worked
instance, see examples/identity-mini-guide.md.
The Consistency Gate (quality gate — run before sign-off)
A distinctive identity that is applied inconsistently reads as accidental, not authored.
The full checklist lives in references/brand-consistency-gate.md (this folder); the summary:
- Logo: correct lockup, clear-space, and minimum size on every surface; single-colour and
dark-mode variants exist and are legible.
- Colour: roles honoured; the signature colour is actually carried through; no off-system
gradient or stray accent crept in.
- Type: the identity pairing is used; no slop "escape" font substituted downstream.
- Imagery: every photo/illustration passes the AI-slop visual checklist
(
doctrine/references/ai-slop-taxonomy.md) — reject and remake any that don't.
- The result passes the memory test: could someone redraw the essence of this identity
after looking away? If not, it is not yet ownable — sharpen one element.
If any item fails, fix before sign-off — do not ship a half-applied system.
Anti-Patterns
- Stock-gradient logo / "AI badge" aesthetic used as identity instead of a real mark.
- AI-generated wordmarks with warped, fused, or misspelled letterforms.
- Interchangeable startup sameness: geometric-sans wordmark + purple-blue gradient +
generic 3D-blob hero — swap the name and nothing changes. That is the convergent mean the
doctrine exists to defeat.
- Designing a logo with no system around it (no clear-space, no min size, no mono variant).
- Citing an AI tool's "brand suggestions" as authority. Approvals trace only to human design
authority; an AI's own picks are exactly what the next wave converges on (doctrine §2,
sourcing-authority asymmetry).
- Five hedged "safe" choices instead of one strong, defensible one (doctrine §0).
Outputs
| Output | Consumer | Evidence / acceptance |
|---|
| Identity-system specification | Designers and partners | Mark, colour, type, layout, imagery, and voice rules |
| Application exemplars | Product, marketing, and document teams | At least three representative contexts |
| Consistency-gate record | Brand owner | Recognition, coherence, accessibility, and misuse checks |
Quality Standards
- Tie every visual rule to the brand promise, audience, or constraint.
- Prove coherence across real applications rather than a mood board alone.
- Stop sign-off when assets are unlicensed, accessibility fails, or differentiation is absent.
Decision Rules
| Condition | Decision | Wrong-choice failure |
|---|
| Existing asset has recognised equity and remains fit | Evolve with documented continuity | Refresh destroys recognition |
| Several directions are plausible | Select against differentiation and application proof | Trend or vote decides taste |
| High-stakes sector needs legitimacy | Apply trust-architecture evidence | Expression weakens credibility |
| Rule cannot reproduce across channels | Simplify or govern by context | Identity fragments in delivery |
Capability Contract
Read and cross-channel inspection are required. Editing and publication require identity authority. Do not alter protected marks, approve licences, or claim market differentiation without evidence.
Degraded Mode
If required evidence or tooling is unavailable, use the scoped fallback below and mark the result unverified.
Without strategy or application contexts, produce a discovery gap list and provisional direction only. Without renderable assets, mark validation conditional.
A mini style-guide (one to three pages) containing:
- Brand essence + adjective set + the not-to-be-confused-with list.
- Logo: lockups, clear-space, minimum size, colour/mono/dark variants, misuse examples.
- Colour roles + signature colour (deferring construction to
color-system-and-palette).
- Type roles (deferring selection to
font-selection-and-pairing).
- Photography / illustration / iconography style rules.
- The voice-of-the-visuals paragraph + one do/don't pair.
- A completed Consistency Gate checklist.
Examples
examples/identity-mini-guide.md — a full worked mini brand identity ("Sonda"): strategy,
logo system, colour tie-in, type tie-in, supporting visual language, voice-of-the-visuals,
an applied one-screen style-guide, and a filled Consistency Gate. Mirror its shape, not its
values.
References
references/identity-system-spec.md (this folder) — the canonical structure every layer of
the identity system must fill; read it before producing the mini style-guide.
references/brand-consistency-gate.md (this folder) — the sign-off checklist run in step 7.
references/trust-architecture-checklist.md (this folder) — when the brand surfaces must
carry proof/legitimacy (legal, consulting, healthcare, financial, high-stakes sectors).
doctrine/design-doctrine.md — Mission (§0, the moat is looking human-made), the Anti-Slop
Charter (§2), and the sourcing-authority asymmetry rule (human design authority only).
doctrine/references/ai-slop-taxonomy.md — corporate/advertising slop (misspelled logos,
uncanny assets) and the visual tells checklist.
- Sibling skills:
../color-system-and-palette (palette construction & contrast) and
../../01-typography-and-fonts/font-selection-and-pairing (typeface selection & pairing);
../../03-layout-grid-and-composition/layout-grid-and-spacing (grid/spacing); ../../00-cross-cutting-ops-qa-a11y/visual-product-slop-audit.
- Human design authority grounding the approach: Paul Rand (a mark is simple, distinctive,
reproducible, and earns meaning through use); Massimo Vignelli (discipline — a small governed
set of elements, used consistently); Marty Neumeier, The Brand Gap (differentiation first;
a brand is what they say it is). No AI-tool recommendation is cited as authority.