The lead-scoring automation we ship in week one

New clients always ask the same question: what do you actually build in the first 30 days?

Here's one thing — the lead-scoring automation that goes live in week one.

The pipeline, end to end

1. Lead form fires (webhook from your site).
2. n8n picks it up, parses fields, normalizes the email domain.
3. Enrichment call to Clearbit / Apollo for firmographic data (headcount, industry, revenue band).
4. AI agent (Claude / GPT) scores on a 0–100 rubric using your historical closed-won data.
5. CRM write-back — HubSpot / Salesforce gets the score, enriched fields, and a reasoning string.
6. Slack ping to #sales for any lead scoring ≥ 70, with a one-line "why" and a CRM link.
7. Re-score nightly as new behavioral signals come in (pages visited, emails opened).

Every step is observable. When a score moves, there's a reason attached. When your sales team disagrees, their override is a signal that feeds the next tuning pass.

What we weight

The rubric is tuned per client, but the universal signals are:

• Firmographic fit — headcount, industry, and revenue band matched against your closed-won profile
• Intent depth — how many high-intent pages the prospect visited pre-form
• Budget alignment — if they picked a budget band in the form, does it match what your pricing supports
• Timing markers — phrases in the message ("urgent", "by Q3", "evaluating") that correlate with 30-day closes

Tuning over time

The first version is a hand-built rubric. Within 4 weeks we have enough labeled data to retrain weights. Within 8 weeks the model has a score-to-close-rate calibration curve you can defend.

This is what we mean when we say "automation, not dashboard." The scoring doesn't live in a UI — it lives in the workflow. By the time a lead is in front of a rep, the prep work is done.

Ship it

Every B2B team should have this running. If yours doesn't, we can have the v1 deployed against your stack in 5 working days. Talk to us.