DataBahn · AI workflows and enrichment
Building a multi-stage Clay workflow for account enrichment and contact sourcing
I built this Clay workbook for two related jobs: enrich target accounts deeply enough to know which companies were worth pursuing, then source and classify the right people before writing outbound. The workflow starts with 150 companies, expands into account and technology signals, routes contacts by seniority and fit, and writes personalized outbound email sequences based on verified account and contact research from the earlier stages.
- Company
- DataBahn
- Role
- Marketing Manager
- Use cases
- Account enrichment and contact sourcing
- Scale
- 4 connected stages, 150 accounts, and 30K+ contact rows
The problem
Raw account lists were not enough. We needed verified technology context, relevant people, valid work emails, clear relationship and topic signals, and outbound that could explain why each account was being contacted.
What I owned
Workflow architecture, enrichment waterfalls, prompt logic, classification rules, email verification, quality checks, and personalized outbound email sequences based on verified account and contact research across the four connected stages.
Key decisions
- Used the full configuration behind “Verify technology usage” to build a multi-provider technology waterfall instead of trusting one database.
- Validated “Work Email” before a contact could move into outreach.
- Used “Relationship Classification” to separate real relevance from superficial title matching.
- Added a clear “What does NOT match” section to “Topic Signal Analysis” so unrelated activity did not become a false signal.
- Built “Write Outbound Email Sequence” to produce personalized outbound email sequences based on verified account, role, relationship, and topic research.
What shipped
- Stage 0: CSV input and normalized company records
- Stage 1: account enrichment and technology verification
- Stage 2: contact enrichment, work email verification, relationship classification, and topic signal analysis
- Stage 3: C-suite and VP enrichment with personalized outbound email sequences based on verified account and contact research
- Quality controls for positive matches, negative matches, missing data, and unsupported claims
Results
- Use case 1: Account enrichment
- Use case 2: Contact sourcing
- 4 connected stages
- 30K+ contact rows processed
Evidence and artifacts
Inside the enrichment stage
What I learned
The useful part of an enrichment system is not the number of columns. It is the logic that decides what counts as evidence, what gets rejected, and what is allowed to reach the message.