content-intelligence — a quality gate before every publish.
Measures AI readiness, brand-voice consistency and freshness of every piece of content — with a backend dashboard, an audit trail for AI-generated content and an extensible analyser pipeline. Anyone producing content sees immediately whether it is ready to publish.
Content quality is usually a gut feeling.
With content-intelligence
- AI readiness score (0–100) per page and per content element
- Brand-voice consistency check against a configurable voice profile
- Freshness analysis with quarterly and yearly thresholds
- Audit trail for AI content (model, prompt hash, reviewer)
- Compliance dashboard in the backend
- Extensible analyser pipeline — register your own checks
Until now
- Brand voice is a PDF nobody has at hand
- AI readiness is a buzzword, not a measurement
- Stale content gets discovered by accident
- No quality gates before publish
- No audit trail for AI-generated content
Four building blocks
Extensible analyser pipeline
Custom checks registerable via AnalyzerInterface — e.g. industry-specific compliance rules or your own brand-voice heuristics.
Compliance dashboard
Backend module with an overview of all content, score distribution, ageing analysis, prioritised action list.
Brand-voice consistency
Compares tone and vocabulary against a configurable voice profile — with hints on what doesn't fit.
AI readiness score
Rates how well a piece of content can be processed by AI systems — structured data, semantic markers, clarity, source chain.
Architecture
Classes/
├── Domain/ # Analysis models, scoring value objects, contracts
├── Application/ # Analysis orchestration services
├── Infrastructure/ # Analyzers, persistence, HTTP clients
└── Presentation/ # Dashboard controllers, event listeners
Database tables
| Table | Purpose |
|---|---|
tx_contentintelligence_analysis_result | Analysis results and scores |
tx_contentintelligence_audit_entry | Audit trail entries |
pages (extended) | Additional fields for content intelligence metadata |
tt_content (extended) | Additional fields for content intelligence metadata |
Configuration and requirements
Configure analysis rules and thresholds in the TYPO3 backend. Brand voice definitions form the baseline for consistency checks. Scoring weights can be tuned per content type.
Requirements
- PHP 8.2+
- TYPO3 14.0
koh/ai-workflows(mandatory — workflow engine)
Dependencies
| Package | Type | Purpose |
|---|---|---|
koh/ai-workflows | Required | Workflow orchestration |
koh/dev | Dev | Shared QA tooling |
Source code & docs
Composer package
composer require koh/content-intelligence – on Packagist, source on GitHub.
Mirror
Development happens on a private GitLab instance. GitHub is the public mirror: every release appears there with its tag, and Packagist reads it from there.
Brand voice and quality gates as a service?
content-intelligence is open source. If you want to run quality gates as part of an editorial setup — including voice profile, thresholds and workflow integration — get in touch.
Or email me directly: mail@ole-hartwig.eu