Prerequisites: finish Build your first workflow, then skim Records vs list entries and Conditions & filters.By the end you can: ingest a lead by webhook, then enrich only its empty fields cheapest-first. You can also score the lead into an A/B/C band that decides the next action.
The mental model
Qualification is a stack of three questions. Fit asks whether this is the right kind of company. Intent asks whether they are showing buying signals. Engagement asks whether they have touched you yet. You check the cheap, free signals first. Then you spend money only on the records that clear the bar. That money goes to enrichment calls and AI classification. This is the signal waterfall: enrich when a signal fires, not on every raw lead. The cheapest filter is a Filter block. It stops a run silently and costs nothing, so it belongs in front of every step that costs credits. The engine vocabulary used across this path lives in the Glossary.Build the three modules in order
Each step is a complete, follow-along build in Guides that you ship into your own workspace. Work them in order. The webhook gives you a clean record. The waterfall fills it in. The score then turns it into a routing decision.1
Take in leads from anywhere, safely
Build a single webhook front door that never creates a duplicate. The Webhook received trigger accepts form fills, product sign-ups, and partner feeds at one endpoint. You Parse JSON field by field, which turns the raw payload into usable values. You drop anything without a usable identity. You then match on a stable key (such as an email or domain) with Find records. You write with Create or update record. That last block means a repeat submission updates the existing record instead of cloning it.Follow the full build in Webhook lead intake with safe upsert.
2
Fill the gaps cheaply
Enrich the record cheapest-first and stop on the first hit. You call providers in price order. You write only the fields that are empty, and you never overwrite verified data. So a record the first provider covers never pays for the second. This is the signal waterfall applied to enrichment.Follow the full build in Cheapest-first enrichment waterfall.
3
Turn signals into a score a rep trusts
Combine fit, intent, and engagement into one transparent score with a Formula block. Then branch on the band with a Switch. The score is plain arithmetic, not the hidden output of an AI model. So a rep can open the run and see exactly why a lead is an A. Each band routes to a different next action.Follow the full build in Score leads into explainable bands.