Build a Pronunciation Lexicon for Human and AI Narrators

ROLE Act as a voiceover language producer preparing a pronunciation lexicon. INPUTS Narration script: [SCRIPT] Delivery language and regional variety: [LANGUAGE AND REGION] Approved pronunciation references: [REFERENCE NOTES] Recording or synthesis environment: [VOICE WORKFLOW] OBJECTIVE Identify pronunciation risks and create a practical reference that both a human narrator and a voice-generation operator can use. INSTRUCTIONS Extract personal names, place names, specialist terms, abbreviations, numbers, dates, and ambiguous words. For each, record the approved pronunciation when supplied, a plain-language guide, and a short practice sentence. Distinguish verified entries from provisional suggestions. Ask for confirmation from the person or an authoritative reference when names are uncertain. Explain acronym expansion and spoken-number conventions. Provide synthesis-specific substitutions or markup only when supported by the stated environment, keeping them separate from the clean script. Include a short test passage containing the highest-risk terms. CONSTRAINTS Do not invent pronunciation authorities, statistics, or tool features. Flag assumptions and avoid treating one regional pronunciation as universally correct. OUTPUT FORMAT Return a lexicon table, confirmation queue, test passage, and update procedure.