An automated B2B prospecting pipeline targeting 103,000 companies, delivered in ten days
Proving that a large-scale prospecting system can be delivered quickly and without heavy development, with a no-code/low-code architecture suited to a tight budget and deadline.
- Client
- A search fund based in Geneva
- Role
- Automation architect & developer
The context
A Geneva search fund needed to identify and qualify acquisition targets among the canton's SMEs, with no budget for heavy development and no time to hire a data team.
What was done
- 1
Building the target database
Extraction and cleaning of a database of 103,000 Geneva companies from SITG public data, filtered by NOGA code to target the relevant sectors.
- 2
Multi-source enrichment
A pipeline combining Zefix (the Swiss commercial register), the Geneva commercial register API through a Swiss residential proxy, Google Places and LinkedIn, to complete each company record with reliable data.
- 3
AI scoring and summaries
Kimi K2.5 used to generate company summaries and personalised outreach messages, replacing manual qualification work that would be impossible at this scale.
- 4
Orchestration and CRM
End-to-end automation with n8n, data centralised in Teable then Attio (CRM), outreach run through Lemlist. Everything connected by an architecture designed to stay maintainable without an in-house technical team.
Tech stack
- n8n
- Teable
- Attio (CRM)
- Lemlist
- Zefix
- Google Places
- Kimi K2.5
What it changes
A large-scale prospecting system delivered quickly, without sacrificing either target coverage or enrichment quality.
Concrete results
Pipeline in production, able to process a database of 103,000 companies, the most complex automation project handled by the agency.