An AI-powered pipeline that discovers key decision-makers at target companies — without paid contact databases. Finds the right people, optionally enriches with LinkedIn and email, and checks your CRM so you never chase someone already in pipeline.
Built as a generalizable n8n workflow. Swap the company list and industry context, and the same pipeline maps the org chart for any B2B target — no Zoominfo, no Apollo, no per-seat licensing.
Three phases, each building on the last. The pipeline reads a company list and outputs a verified, deduped contact map ready for outreach.
The pipeline starts with a filtered list of top accounts from your CRM — the companies your team has already identified as high-priority targets. For each one, it runs structured web research to identify the people who matter. It doesn't just return a list of names — it maps the org chart by function: who runs revenue, who owns customer success, who leads product. The AI distinguishes between decision-makers and influencers, and captures context about why each person is relevant to the outreach motion.
The pipeline has a modular enrichment node. Plug in any contact data provider — Apollo, Hunter.io, Lusha, or a custom API — and it appends LinkedIn profile URLs, verified work emails, and phone numbers where available. The enrichment is optional: the core discovery phase works without any paid data source, and the enrichment layer only fires when you choose to connect one.
Provider-agnostic design means you're not locked into any single vendor. Swap APIs by changing one credential, not the pipeline.
Before any contact lands in your outreach list, the pipeline checks your CRM. It matches by email domain, name similarity, and company — flagging contacts that already exist, are in active sequences, or are associated with open deals. The output separates "new" from "existing" so you never burn a sequence on someone already in pipeline.
The pipeline architecture is constant. What changes per deployment is a handful of prompts and configuration values — making it reusable across any company, industry, or GTM motion.
| Layer | What Stays Universal |
|---|---|
| Company Input | Filtered CRM account list → batch processing pattern |
| Org Research | Structured web search + LLM extraction |
| Enrichment | Modular provider node with standardized output schema |
| CRM Dedup | Domain/name/company matching logic |
| Output | Structured contact record format |
Most GTM teams pay $100–400/seat/month for contact databases. This pipeline does the core discovery work with AI web research — free. Enrichment is optional and provider-agnostic.
The dedup phase means you never sequence someone already in pipeline. No more awkward "I know we're already talking but..." emails. The pipeline respects your existing relationships.
Once the pipeline exists, deploying it for a new company or industry is a prompt-edit exercise. Change the research context, adjust role definitions, point it at a new sheet — done.
This pipeline represents a philosophy: GTM infrastructure should be built once and reconfigured with prompts, not rebuilt from scratch for every new market. The architecture is the asset. The prompts are the configuration.