A Sector-Specific Lead Generation Engine
We built lead generation and outreach automation across four sectors. The core engine is the same each time: find, enrich, write personally, send on schedule, classify the reply, follow up. What made the difference was not the engine but rebuilding it around each sector's regulation and each company's tooling.
4
Sector-specific builds
28,852
Rows processed without an LLM
191
Dormant targets found automatically
The problem
Four companies in four different sectors had the same underlying need: find the right contact, write them something meaningful, and manage the reply without losing it. On the surface it looked like one solution. In practice each had different data sources, email infrastructure, data protection obligations, and sales language — and since all four operated in Germany, DSGVO compliance was mandatory across the board.
How we approached it
We held the core engine constant and layered company-specific architecture decisions on top of each build. We did not copy and paste the same system, because the sector's regulation and the company's tooling changed how the core had to behave in the first place. Positive replies never sending automatically in healthcare, the system being built on Microsoft infrastructure in real estate, LLM-free data processing for cost control in water treatment — none of these are settings, they are separate design decisions.
Builds that diverged by sector
The core engine is identical in all four: find, enrich, write personally, send on schedule, classify the reply, follow up. The differences below were written into that engine, not bolted onto it.
B2B consulting and training
Germany · GDPR deletion flow
- A four-stage pipeline with a separate AI prompt at every stage
- Missing email addresses extracted automatically from a company's Impressum page
- GDPR deletion flow included from the start, with a three-month follow-up cycle
Real estate investment
Germany · Microsoft 365
- Built from scratch on Azure AD and the Microsoft Graph API
- Excel lead database, a shared Outlook mailbox, and OneDrive reporting
- A token-limiting node for large sites, with full DSGVO compliance
Healthcare and medical equipment
Germany · High sensitivity
- Positive replies never send automatically: the system drafts, a human approves
- Clinic scraping via Google Maps alongside web and social media analysis
- DSGVO-compliant soft delete: content is removed, the audit log is kept
Industrial water treatment
Germany · Three segments
- New customers, dormant-customer reactivation, and portal registrations on one shared backbone
- Last-purchase data pulled from a 28,852-row Excel file via in-memory JOIN, with no LLM involved
- 191 dormant targets segmented automatically out of 875 customers
Outcome
- Four separate builds across four sectors on one shared core engine
- Four-way reply classification standard in every version
- GDPR and DSGVO compliance designed in rather than layered on, in all four
- Active error handling in place in every version
Stack
What this project generalises to
These four projects show that what scales in automation is the method, not the product. We reused the core engine four times but made the architecture decisions afresh each time. Had we deployed the same system four times, we would have deployed the wrong system four times.
The service this case is evidence for
n8n Workflow AutomationOther case studies
- A 24/7 Multilingual WhatsApp Sales Agent
- RAG-Based Support Automation in Healthcare Training
- Five Parallel AI Workflows in Regulated Manufacturing
Client names are withheld under confidentiality; the sector and region describe the real engagement.
Last updated: 7 August 2026
Talk About Your Own Sector