mirror of
https://github.com/domfelipe/vibeflow-n8n.git
synced 2026-08-07 07:56:42 +00:00
48 lines
1.3 KiB
Markdown
48 lines
1.3 KiB
Markdown
# Recipe: AI Lead Enrichment
|
|
|
|
## Goal
|
|
|
|
Capture a new inbound lead, enrich the company context, generate a short structured summary, score the lead using simple rules, and write the result to the CRM.
|
|
|
|
## Typical trigger
|
|
|
|
- new CRM lead
|
|
- form submission
|
|
- webhook from a landing page
|
|
|
|
## Systems involved
|
|
|
|
- CRM such as HubSpot or Pipedrive
|
|
- enrichment source such as Clearbit-like data or internal lookup
|
|
- optional LLM step for summarization
|
|
- Slack for alerts
|
|
|
|
## Core steps
|
|
|
|
1. Receive the lead payload.
|
|
2. Normalize fields such as company name, email domain, and source.
|
|
3. Query enrichment data.
|
|
4. Generate a concise structured summary.
|
|
5. Apply a scoring rule based on company size, geography, and source.
|
|
6. Update the CRM record.
|
|
7. Notify Slack if the lead crosses a threshold.
|
|
|
|
## Important assumptions
|
|
|
|
- scoring starts with safe defaults
|
|
- enrichment can fail without blocking the whole workflow
|
|
- missing optional fields should not break CRM updates
|
|
|
|
## What makes this recipe good for demos
|
|
|
|
- shows practical AI usage
|
|
- stays grounded in business logic
|
|
- produces an output people immediately understand
|
|
|
|
## Validation checklist
|
|
|
|
- lead payload is parsed correctly
|
|
- enrichment failures are captured clearly
|
|
- summary format is stable
|
|
- score thresholds are visible and editable
|
|
- CRM update succeeds even when optional data is missing
|