Automatically enriches leads, researches individuals and companies, scores against ICP, and routes hot, warm, and cold leads with notifications and CRM updates.
The AI agent receives a lead via webhook, enriches the contact with firmographic data from PDL, and researches the individual's activity and company developments with Perplexity. It optionally scrapes LinkedIn data through Apify to augment the profile. Claude then scores the lead against ICP rules stored in Google Docs, and routes the lead based on the score to Slack alerts, digest channels, or CRM updates.
Concrete actions that automate qualification
Ingests leads via webhook with email and optional name
Enriches data with PDL (contact and firmographic details)
Researches individual activity and company developments with Perplexity
Optionally scrapes LinkedIn profiles via Apify to augment data
Scores leads against ICP rules using Claude AI
Routes hot leads to Slack with email drafts, warm leads to a digest, and cold leads to CRM
Before an AI agent, manual lead research is slow and data quality is inconsistent; after, enrichment, scoring, and routing are automatic with real-time alerts and CRM updates.
A simple 3-step AI agent flow
Receives a webhook with email and optional name and starts parallel enrichment processes.
Aggregates data from PDL, Perplexity, and optional LinkedIn, then scores the lead against ICP rules stored in Google Docs using Claude.
Routes hot leads to Slack with email drafts, warm leads to a digest channel, and cold leads to the CRM; optionally syncs data back to your CRM.
A realistic scenario of processing a lead batch
Scenario: 40 new inbound leads arrive via webhook. Each lead is enriched in ~60 seconds with PDL and Perplexity, optionally augmented by LinkedIn data. Claude scores each lead against ICP rules; 6 hot leads trigger instant Slack alerts and generate personalized email drafts, 14 warm leads are posted to a digest channel, and 20 cold leads are logged to the CRM for follow-up.
Roles that gain speed and clarity in lead qualification
Spend less time researching and more time outreach.
See real-time funnel health and scoring distribution.
Centralize enrichment and scoring across teams.
Improve data quality for attribution and routing.
Prioritize outreach to hot and near-hot leads.
Auto-sync enriched profiles into the CRM.
Connects data, AI scoring, and channels
Enriches contact and firmographic data for leads.
Researches individual's activity and company developments.
Scores leads against ICP rules using Claude.
Stores ICP scoring rules used by Claude.
Sends hot lead alerts and routes warm leads.
Sends personalized email drafts for hot leads.
Optionally augments profiles with LinkedIn data.
Upserts or syncs qualified leads to the CRM.
Practical scenarios to apply this AI agent
Common questions about using this AI agent
Typically under a minute per lead when all data sources respond promptly; in practice, the total processing time is 30-60 seconds per lead. The parallel enrichment steps (PDL, Perplexity, and optional LinkedIn) run simultaneously to minimize wait time. If any source is slow, the system continues with available data and updates the score as new signals arrive. You can tune the scoring thresholds in the ICP rules document to reflect your preference for hot/warm/cold definitions.
Yes. You can replace PDL with Apollo or Clearbit via HTTP requests and adjust the Merge Enrichment step to parse the new response. The scoring logic in Claude remains the same; you would map new fields to the ICP criteria. Ensure the new data format provides the required fields (company size, industry, title, etc.). You may need to update any mappings in the Merge All Sources step. Tests are recommended after swapping sources.
LinkedIn scraping is optional. If you disable the LinkedIn scraping node, the workflow still enriches with PDL and Perplexity and uses Google Docs ICP rules for scoring. You can remove this step to simplify for n8n Cloud compatibility. When added, scraping respects user consent and privacy considerations per your policy. The rest of the enrichment and scoring pipeline remains unaffected.
ICP rules are stored in a Google Doc referenced by the ICP & Use Case node. This keeps scoring criteria human-readable and easy to update without redeploying the AI agent. The agent reads the document in real-time to calculate points for company fit, title fit, buying signals, and timing. You can adjust the document to reflect changing ICP definitions without touching the code.
Yes. The workflow can Upsert or Sync leads to HubSpot, Salesforce, Pipedrive, or other CRMs. CRM integration is optional and configurable per your pipeline. If the CRM is unavailable, the agent still routes notifications and stores data via logs, enabling later sync. Real-time updates can be enabled or disabled based on your data governance needs.
The AI agent processes data according to the sources you configure. Ensure you have consent and comply with data protection regulations. Secrets and API keys are stored in your credential manager and never hard-coded. You should implement access controls and audit logs for data enrichment and storage actions.
Scoring thresholds live in the ICP rules Google Doc. Edit the ranges for hot/warm/cold and add new factors like tech stack or budget authority if needed. After updating, re-run test leads to validate the impact of the changes. The AI agent automatically uses the latest version of the rules document on next runs.
Automatically enriches leads, researches individuals and companies, scores against ICP, and routes hot, warm, and cold leads with notifications and CRM updates.
