es-ES7 / 8 approved
Estrena Creator Pro
Lleva tus mezclas más lejos: nuevos efectos, exportación en alta resolución y una biblioteca de samples lista para producir desde el primer compás.
1 term needs reviewUse case · Generative marketing
Marketing campaigns written for the market, not translated into it.
Knowing your market is knowing how to write for it. A plan in Germany is not a plan in Mexico. Black Ice governs the concept before the copy — so AI-generated campaigns are built on what each market understands, not on what the English brief says.
The problem
Translation flattens the argument.
Every market sells a different product — different price points, different compliance constraints, a different definition of "benefit." Running an English brief through a multilingual AI generates text in twelve languages and a single argument. Black Ice intercepts at the concept layer.
Without Black Ice
- AI translates the English brief as-is
- "Free" framing that works in the US reads as suspicious in France
- Compliance-sensitive terms cross market lines without governance
- Hallucinated terminology has no approval record
- Localization QA catches errors — after copy is written
With Black Ice
- Brief by outcome, not by source string
- Approved concept per market governs generation
- Per-market disclosure logic lives at the concept level
- Every term has an ontology status — approved, planned, flagged
- Governance runs before copy — not after
How it works
Concept governance before copy generation.
Black Ice sits between the brief and the generative model. The pipeline cannot produce a term that hasn't cleared the ontology.
- 01input
Brief by outcome
The team submits a market-neutral outcome — "Spring · No-fee card" — not a translated source string. No language is privileged.
- 02list_concepts
Live ontology resolves per-market concepts
Black Ice pulls the approved terminology for each target market. Each concept carries governance metadata: status, allowed contexts, compliance flags.
- 03generation
Copy generated within approved concept space
The AI writes market-native copy constrained by the ontology. It cannot invent a term that hasn't been approved — the governance layer is structural, not advisory.
- 04validate_translation
Batch validation before human review
The full generated batch runs through validate_translation. Approved terms surface immediately. Anything without a sanctioned equivalent is flagged — not silently passed through.
- 05check_term · flagged
Flagged terms halt launch copy
Unresolved concepts — terms where fr-FR has no approved equivalent, or de-DE has a suggested but unconfirmed term — are surfaced with context and a recommended action. The pipeline stops. A human decides.
Sample output
What the pipeline produces.
For a product launch brief across three markets — es-ES, fr-FR, de-DE — Black Ice governs concept approval per market and halts launch copy when terms are unresolved.
fr-FR8 / 8 approved
Creator Pro arrive
Passez au niveau supérieur : mixage haute définition, exports sans compromis et une bibliothèque de samples pensée pour les productions exigeantes.
All approvedde-DE6 / 8 approved
Jetzt neu: Creator Pro
Mehr Kontrolle im Mix, verlustfreie Exporte und eine kuratierte Sample-Bibliothek — entwickelt für Producer, die professionelle Ergebnisse erwarten.
2 terms need reviewYou need this if
Four signals this problem is yours.
- You brief by outcome, not by source string
- Your markets require different arguments, not just different words
- A "free" framing that works in the US reads as suspicious in France
- You're in regulated industries where every market variant needs its own disclosure logic
Feature · Market Definitions
Brand voice isn't translated.
It's defined — per market.
Writing that works starts before the first word: knowing who you're addressing, how you want to speak to them, and why. Each locale in Black Ice carries a Market Definition — ten default sections, customisable per workspace, that record exactly that, per market. Not just what words to use, but why the register is different, what cultural expectations are implicit, and what is never allowed. So AI agents and human writers don't guess.
CULTURE DNA PROFILEde-DE · German market
Read the docs →Voice & Locale DNA
Formal, precise, understated. Brand keywords: Zuverlässigkeit, Qualität. Anti-patterns: superlatives, exclamation marks, casual second-person.
Tone & Style Rules
Formal register. Third-person address where expected. No contractions. Prefer passive constructions for feature descriptions.
Culture & Pragmatics
"Free" framing requires explicit qualification. Price transparency is non-negotiable. Understatement outperforms enthusiasm.
Output Constraints
Never claim "kostenlos" without legal qualifier. Always include Impressum reference in footer copy. No US-style testimonial formats.
Golden Examples
"Exportieren Sie Ihre Projekte verlustfrei — in jedem Format." Not: "Export anything, anywhere, for free!"
Locale Conventions
Decimal separator: comma. Date: DD.MM.YYYY. Currency: 9,99 € (symbol after). Pricing psychology: even numbers preferred.
Each Market Definition covers ten structured dimensions — available to every AI agent in your pipeline over MCP, the workspace API, or as structured Markdown synced to GitHub. Same rules for your models and your human writers.
Market SnapshotVoice & Locale DNATone & Style RulesCulture & PragmaticsGrammar & MechanicsUX Microcopy RulesLocale ConventionsTerminology RulesGolden ExamplesOutput Constraints
Concept-first authoring
A single concept carries the goal, the compliance rules, and the market policies. Copy is downstream.
Native per-market variants
Each market gets copy generated against its own price, tier, tone, and disclosure rules — no translation tax.
Auditable governance
Every variant ships with a trace of which policies fired, approved by whom, and at which version.
FAQ
Common questions about generative marketing.
Stop translating English into 14 broken campaigns.
Govern the concept once. Generate market-native campaigns on demand, with the audit trail your legal and brand teams need.