Zoomlion Duqm Lead Generation

Zoomlion needed a repeatable way to find heavy-equipment buyer prospects tied to the Duqm Special Economic Zone — EPC contractors, civil engineering firms, quarry operators, and equipment rental companies operating in Oman. Manual Google searches produced inconsistent results, duplicate companies, and no structured evidence that a prospect was actually Duqm-relevant. Enrichment steps (commercial registration numbers, contact emails, website validation) lived in separate spreadsheets and ad-hoc browser tabs, making each research run slow and error-prone.

Built a 68-node n8n workflow that orchestrates search, AI extraction, multi-stage qualification, and sheet output end-to-end.

1. Config-driven ICP — Campaign search templates and Duqm keywords load from a Google Sheets config tab so operators can tune queries without editing the workflow.
2. Serper-powered discovery — Targeted domain searches collect candidate companies; deduplication runs against existing sheet rows and a domain blocklist filters aggregators and news sites.
3. Gemini extraction — Each search result page is fetched, stripped to text, and passed to Google Gemini to extract company name, website, industry guess, Duqm evidence, and a Zoomlion relevance score.
4. Qualification gates — IF nodes enforce Duqm verification, commercial registration lookup (website scrape + Serper CR/PDF searches), Hunter email enrichment, and entity exclusion rules before a lead is accepted.
5. Batch orchestration — SplitInBatches nodes process candidates with static workflow data tracking accepted/rejected counts and rejection reasons across the full run.
6. Sheet output — Qualified companies append to the tab with a summarize node reporting run statistics.

• Fully automated lead pipeline — One manual trigger runs the entire discovery-to-sheet flow across dozens of search queries.
• AI-structured extraction — Gemini normalizes messy HTML search results into consistent JSON fields with Duqm evidence and buyer-fit scoring.
• Multi-source CR enrichment — Website scraping plus Serper CR and PDF searches improve Oman commercial registration coverage before acceptance.
• Built-in deduplication — Domain and URL checks against existing sheet rows prevent duplicate company entries.
• Observable rejection tracking — Per-run counters for not_duqm, no_cr, parse_error, excluded_entity, and non_oman_vendor make tuning the ICP measurable.
• Operator-friendly config — Search templates and keywords live in Google Sheets, not hardcoded in the workflow.