Surface what standard QA was never designed to catch.
Point Black Ice at content you've already published — locale files, translated pages, help content, chatbot training data. It scans against the live ontology and surfaces terminology drift, missing disclosures, restricted feature mentions, and tone inconsistency across every market.
| Defect type | Severity | Markets affected | Count |
|---|---|---|---|
| Terminology drift | critical | de-DE · fr-FR · ja-JP | 6 |
| Disclosure gaps | critical | de-DE · fr-FR | 2 |
| Restricted phrasing | high | es-ES | 1 |
| Tone inconsistency | medium | de-DE · fr-FR · pt-BR | 3 |
What Black Ice finds that your TMS never flagged.
Standard QA checked translation fidelity at the moment of translation. It had no way to know what changed in the ontology after the content shipped. Black Ice does.
Terminology drift
A term was correct when it was translated — but the concept has since been renamed, deprecated, or reassigned in the ontology. The old term is still live in published content. Standard QA has no way to detect this because it checked fidelity to the source, not currency against the ontology.
Disclosure gaps
A governed concept is referenced in a market where a specific regulatory disclosure is required — but the disclosure is absent from the content. The content may have been translated before the disclosure requirement was added to the ontology, or the governance rule may have changed after publication.
Restricted phrasing
A formulation that is forbidden in the market — legally, competitively, or by brand mandate — is present in published content. Forbidden variants are defined at the concept level in the ontology. This defect type finds cases where they were used anyway, either before the restriction was added or because the rule wasn't applied consistently downstream.
Tone inconsistency
The content diverges from the register, formality level, or voice conventions defined in the market's profile. This is not a terminology error — the words may be correct — but the overall register is wrong for that market. Common in content translated before market definitions were formalised.
Restricted feature mentions
A feature or plan referenced in the content has since been gated, removed, or restricted in that market. The translation was correct at the time — but the market availability has changed in the ontology and the content was never updated to reflect it.
The audit is the governance baseline.
You don't need a rebuilt pipeline to start. The audit works against what you already have, in the formats you already use. Most teams use the findings to make the case for rebuilding the pipeline with governance at the source — because the report makes the debt visible.
Start without rebuilding anything
The audit is deliberately the entry point for teams new to semantic governance. No pipeline changes required. Point Black Ice at what you have, get a concrete picture of where governance is absent, and decide what to fix first. The report becomes the brief for the remediation sprint.
Four signals your content needs an audit.
- You've localized without a shared concept layer and suspect terminology has drifted inconsistently across markets since the original translation.
- Features have shipped or been gated since the last translation pass — and no one has systematically updated the content in every affected locale.
- Your legal team has flagged disclosure gaps in published copy — or you're in a regulated market and need to verify compliance before a review.
- You want a governance baseline before rebuilding the localization pipeline — to know the scope of what needs fixing and prioritise accordingly.
Severity-ranked findings
Every defect is categorised, assigned a severity, and linked to the specific string and market where it was found.
Remediation brief
Each finding carries a recommended action — update the term, insert the disclosure, flag for human review, escalate to legal.
Governance baseline
A concrete picture of where your existing content sits against a structured concept layer. The input to the remediation sprint, not just a report.
Common questions about content audits.
Make the governance debt visible.
Book a 30-minute demo. Bring a locale file, a translated page, or a chatbot training set — we'll run a live scan and walk you through the findings.