Voice CV

Voice CV addresses a real problem in mass recruiting: a large share of candidates for blue-collar, logistics, construction, food-service and care roles have no written CV, or have outdated, poorly formatted CVs in a language other than the one required. The result is very low conversion on traditional application forms.

With Voice CV the candidate records an audio message in their own language (often not the local one: Hindi, Arabic, Urdu, Bengali, Punjabi are the five most common languages in blue-collar applications received by European staffing agencies). The system automatically runs: transcription (multilingual ASR), translation, entity extraction (name, year of birth, prior experience, language, role), normalisation onto the ESCO ontology, and generation of a bilingual structured CV.

Typical conversion rates on blue-collar applications: traditional web form 8-12%, simplified mobile form 18-25%, WhatsApp Voice CV link 55-70%. The conversion jump is driven by friction reduction: the candidate doesn't need to type, download an app or format anything.

For recruiters the benefit is twofold. It widens the top funnel (more candidates collected for the same ad spend) and improves CV quality because voice carries soft signals the written CV loses (expressive clarity, language fluency, tone). Generated CVs remain editable by the candidate before entering the pipeline.

The technology demands attention on three fronts: privacy (recordings must be treated as biometric data under GDPR), linguistic bias (some regional accents transcribe worse and require fine-tuning), and candidate confirmation on extracted data to avoid AI hallucination.