How this rule is fixed

This is an AI-enabled rule. Pass/fail requires model judgment; there is no deterministic fixer in the pilot registry.

• Harness: invoke-ai-ruleset-harness.sh runs design-rules/ai/run-design-ai-rule.sh on the Before fixture and expects findings with matching principleId.
• Remediation: Cursor agent plans from forge-ux-remediation.plan.md after sitewide audit — not handbook After copy.

Detection module: docs/design/ux-audit/ai-enabled-design-principles.md#ai-premium-enterprise-feel. Scroll down for Before / After examples and Evidence and remediation steps.

Purpose

Kitchen Sink landing_page, product_page, and forge-card surfaces should read as one deliberate system—calm enough for enterprise buyers, confident enough to trust with governance work. This AI rule judges holistic polish: spacing generosity, typographic refinement, consistent radii and elevation, and imagery that earns its place. It does not replace token or contrast checks; it catches when everything technically passes yet still feels like assembled template fragments.

Deterministic rules guard elevation tokens, font stacks, and grid classes (DET.SURFACE.ELEVATION_TOKEN, DET.TOKEN.NO_DRIFT, DET.LAYOUT.GRID_CONSISTENCY); this rule guards perceived finish on screenshots and DOM review.

Plan: Walk the page top to bottom and note where rhythm breaks—cramped bands, mixed card treatments, ad-hoc inline styles, or decorative layers that cheapen copy. Do: Re-stage sections with KS layout primitives (landing-hero, forge-section, section-label, consistent forge-card variants). Check: First viewport and one mid-page band feel spacious, aligned, and typographically controlled. Adjust: When the same polish failure repeats (for example mixed card padding or competing accent stacks), propose a deterministic DET.* companion or catalog contract tightening.

Passing signals

• Hero uses the landing-hero / fs-landing-hero-band stack with product-landing-title, a single landing-hero-tagline, and a restrained CTA pair—copy bands breathe; nothing feels crammed into the fold.
• Section rhythm repeats forge-section + section-label text-cyan + one h3 heading; vertical padding (py-5, g-3 / g-4) is generous and consistent between bands.
• Cards share one elevation language: forge-card with either default or card-amber, paired with breathe-link or breathe-static—not a mix of inline borders, one-off shadows, and unrelated Bootstrap card classes.
• Typography stays on the Forge scale: font-display for headlines, forge-support for body and hints, card-label for tile eyebrows—no arbitrary inline font sizes or weight stacks.
• Imagery and diagrams are purposeful: product visuals in landing-hero-visual or labeled forge-diagram sections support the claim; decorative noise does not compete with text.
• Accent usage is restrained: amber and cyan appear in predictable roles (primary CTA, section labels, one card accent column)—not every tile shouting a different color.
• Surfaces respect tokenized elevation and radius; dark background, glass cards, and ambient layers feel cohesive with forgesdlc-theme.css / forge-theme.css—not a patchwork of legacy marketing classes.

Failing signals

• Template soup: adjacent sections use different card shells (forge-card, raw glass, Bootstrap card, inline-styled divs) with mismatched padding and corner radii.
• Cramped first screen: hero copy, three outcome tiles, stat band, and announcement strip stack with g-1 / g-2 gaps—reads hectic despite passing numeric burden caps.
• Typography drift: headlines mix h1–h5 without hierarchy, inline font-size overrides, or body copy outside forge-support / landing-hero-tagline patterns.
• Competing accents: every forge-card uses a different accent modifier, gradient text, and button style—no calm focal path (AI.VISUAL.HIERARCHY_CONFIDENCE).
• Decorative filler: stock-style icon grids, placeholder diagrams, or expressive ambient mesh behind dense copy—visuals do not explain product structure (AI.VISUAL.PRODUCT_EXPLANATORY_VALUE).
• Hype stacking: loud gradients, badge rows, and superlatives layered on top of already busy layout—undermines enterprise calm (AI.CREDIBILITY.NO_OVERCLAIM).
• Broken rhythm: columns misaligned across rows, inconsistent container vs container-fluid widths, or sections that jump between full-bleed and narrow rails without reason.

Before example

Before (failing example)

PLATFORM · AI · GOVERNANCE · DOCS

Everything you need!!!

Agents, automation, compliance, analytics, dashboards, and more in one place.

Get started Learn more Book demo Watch video

Fast

Ship instantly

Secure

Enterprise grade

Smart AI

Why teams choose us

Feature

Feature

Feature

Feature

After example

After (passing example)

Forge Platform

Governed delivery you can inspect

Structure intent, delegate safely, and keep human review points—presented with calm, deliberate spacing.

Start with Quickstart See how it works

Outcomes

What changes for your team

Control

Human-owned gates

Review points stay visible—copy and card treatment match the rest of the page.

Evidence

Inspectable artifacts

Logs and contracts linked with the same elevation and typography rhythm as adjacent tiles.

Depth

Mechanism on demand

Technical detail deferred to docs—this band stays spacious and scannable.

Operating model at a glance

Intent

Shape

Clear problem statement before delegation

Execution

Delegate

Bounded jobs and governed LLM tasks

Assurance

Review

Evidence you can trace—not decorative KPI noise

Evidence and remediation

Capture: full-page and first-viewport screenshots; note section spacing (py-*, g-*), card variant mix, typography classes vs inline overrides, accent usage, and whether imagery supports the story. Record relevant hashes (hash="Ldg", outcome section roots) and theme pack if present.

Remediate (in order):

1. Restage the hero on landing-hero / fs-landing-hero-band with one kicker, title, tagline, and CTA pair—remove extra copy bands and button rows that break calm (AI.CONTEXT.BURDEN_SUBJECTIVE).
2. Normalize cards to forge-card + card-amber (when accent is needed) + breathe-link / breathe-static; drop inline shadows, borders, and one-off Bootstrap card shells.
3. Restore section rhythm: forge-section or fs-landing-section, section-label, consistent container / landing-hero-wide widths, and g-3 / g-4 gutters (DET.LAYOUT.GRID_CONSISTENCY).
4. Align typography to font-display, product-landing-title, forge-support, and card-label; remove inline font-size and weight hacks.
5. Tame accents and ambient layers so cyan/amber appear in predictable roles (AI.THEME.PERSONALITY_COHERENCE); keep expressive mesh out of text-heavy heroes.
6. Swap decorative imagery for product-explanatory diagrams or defer dense topology below the fold (AI.VISUAL.PRODUCT_EXPLANATORY_VALUE).
7. Re-run AI batch with principleId: AI.PREMIUM.ENTERPRISE_FEEL; set deterministicCoverage and propose DET.* promotion when the same polish defect repeats (for example banned inline style on hero roots or mixed card padding tiers).

Related rules

• AI.VISUAL.HIERARCHY_CONFIDENCE — scale, contrast, and grouping establish an obvious focal path without equal-weight shouting.
• AI.VISUAL.RHYTHM_SUBJECTIVE — spacing and grouping feel rhythmic, not accidental pile-ups.
• AI.THEME.PERSONALITY_COHERENCE — theme accents and dark/light balance reinforce brand temperament without fighting content.
• AI.CREDIBILITY.NO_OVERCLAIM — hype stacking and fake proof rows undermine enterprise calm.
• AI.CONTEXT.BURDEN_SUBJECTIVE — numeric caps can pass while the page still feels visually noisy and cramped.
• DET.SURFACE.ELEVATION_TOKEN — card and glass surfaces use governed elevation roles, not ad-hoc shadows.
• DET.TOKEN.NO_DRIFT — colors, radii, and spacing stay on Forge tokens rather than inline overrides.
• DET.LAYOUT.GRID_CONSISTENCY — column widths and gutters align across sections on the same page mode.