Case Study - GenAI News Rewriter for Industrial Workers at CCB
Necko Technologies built a GenAI-powered content adaptation tool integrated into CCB's internal mobile app, automatically rewriting corporate SharePoint news into plain, worker-friendly language in French and Dutch.
- Client
- CCB – Compagnie des Ciments Belges
- Year
- Service
- GenAI, Professional Services
Making internal communications reach the shop floor
CCB – Compagnie des Ciments Belges is a Belgian industrial company with over 100 years of history, specialising in the production of cement, aggregates, and ready-mixed concrete. A subsidiary of Cementir Holding, CCB operates the largest cement production facility in the France-Benelux region, several limestone quarries, and 14 concrete mixing plants.
Most of CCB's workforce are on-site employees: quarry operators, plant technicians, logistics staff. These profiles rarely read — let alone to the end — articles written in a formal corporate register. News published on SharePoint by the communications team was simply not reaching the people it was meant for.
- Rewriting time per article
- -85%
- Generated in a single Bedrock call
- FR + NL
An editorial workflow augmented by Amazon Bedrock
Necko integrated a GenAI rewriting module directly into CCB's internal mobile app. When an editor imports an article from SharePoint, a single Amazon Bedrock (Claude Sonnet) invocation generates both a French and a Dutch version, adapted to a worker-friendly register: short sentences, plain vocabulary, direct tone.
The editor reviews, adjusts if needed, and publishes. Content never leaves CCB's environment.
The workflow is intentionally semi-automated: GenAI removes the rewriting effort, not the editorial judgement. For safety communications, a dedicated prompt preserves imperative phrasing even while simplifying vocabulary.
A lightweight architecture, extensible to the Cementir group
The solution relies on a minimal serverless stack: AWS Lambda orchestrates the Bedrock calls, Amazon API Gateway exposes the endpoints to the mobile app, and Amazon DynamoDB stores articles (both original and adapted versions). No new data infrastructure was required.