The Lead Intelligence System is an automated pipeline that identifies, enriches, and prioritizes manufacturing companies in New York State as candidates for energy incentive programs. It turns weeks of manual research into an automated, continuously running system that outputs a scored, sales-ready list.
The Challenge
Finding manufacturers eligible for energy incentives across New York State is a needle-in-a-haystack problem. Companies had to be verified as active NY manufacturers, confirmed to have in-house operations (not third-party-managed facilities), and ranked by estimated utility spend before anyone on the sales team picked up a phone. Doing this manually meant weeks of research per batch with no consistency in data quality.
How It Works
The system runs as a five-stage automated pipeline:
1. Lead Import
CSV files are ingested, field-mapped against a configurable schema, and batch-imported into a structured processing queue. Deduplication, address normalization, and initial classification run automatically before any enrichment begins.
2. Enrichment Orchestration
The main orchestration workflow fetches company data, validates the NY address against a facilities database, confirms in-house manufacturing operations, estimates annual utility spend for electricity and gas, and triggers downstream enrichment stages. Leads that fail qualification are logged and disqualified without consuming further processing.
3. AI-Parallel Research
Three AI agents — GPT-4, Gemini, and Claude — independently research each facility in parallel, each sourcing data from different web endpoints. Their findings are merged and cross-validated to produce a consensus enrichment result. Conflicting signals trigger a validation step before the record is finalized.
4. Contact Enrichment
Apollo.io searches surface decision-maker contacts — Facility Manager, Operations Director, Energy Manager — for each qualified company. Email addresses are verified via ZeroBounce before being written to the database, ensuring the sales team only receives deliverable contacts.
5. Scoring and Output
Each facility is scored on estimated annual utility spend, manufacturing scale indicators, and capture probability. Companies with the highest scores rank first in the output. The result is a prioritized list the sales team can act on immediately — no data prep, no manual research.
The Result
What previously required weeks of manual research per batch now runs automatically. The sales team receives a scored, enriched pipeline of verified NY manufacturers ranked by energy incentive potential — ready to work with no data preparation required.
Built on: n8n · GPT-4o · Gemini · Claude · Apollo.io · ZeroBounce · PostgreSQL