Built-in Enrichments
Alongside enrichment Actions that call an external provider, the New Column search has a set of built-in utility actions. Most of them transform a row’s own data instead — no external API, no credits. There are exceptions: Score Leads judges the row rather than reshaping it, and the two AI Scraping actions read pages off the open web and are billed per page.
They show up in the search modal under AI Scraping, Extract Data, Format Data, and Qualify and Scoring.
AI Scraping
| Action | What it does |
|---|---|
| AI Web Scraper | Extracts whatever you describe in plain English from any URL — 0.4 credits per page |
| Website Contact Scraper | Crawls a site for emails, phones, socials and addresses — 0.2 credits per page |
Extract Data
| Action | What it does |
|---|---|
| Parse JSON | Breaks a raw JSON string into individual fields |
| Merge Columns | Combines two or more columns into one, with a separator |
| Randomizer (Spintax) | Picks a random variation from spintax text per row |
Format Data
| Action | What it does |
|---|---|
| Formula Generator | Turns a plain-English description into a formula column |
| Send To Another Table | Pushes the row into a different table in your workspace |
| Lookup Rows In Table | Matches against another table and pulls back a value |
| Delete Row | Removes the current row, usually behind a filter on Auto-run |
| Separate Name | Splits a full name into First Name and Last Name |
| Get First Name | Pulls just the first name out of a full-name column |
| Get Last Name | Pulls just the last name out of a full-name column |
Qualify and Scoring
| Action | What it does |
|---|---|
| Score Leads | Rates each row against your ICP and writes a score you can filter and gate on |
Common use cases
- Scrape a directory, marketplace or pricing page no built-in scraper covers
- Pull emails, phones and social profiles off a company website
- Break a webhook or import payload into columns you can actually enrich from
- Split or rejoin name fields so personalization tokens resolve
- Vary outreach wording so a large send does not look templated
- Check another table before paying to enrich a row you already have
- Move qualifying rows into an outreach table, and drop the rows that fail
- Score a list against your ICP so paid columns only run on rows worth the credits
Best practices
- The Extract Data and Format Data actions cost no credits — do all reshaping there, before any paid column runs
- The AI Scraping actions are the billed ones in this group: test a single row at one page before raising the page limit
- Keep the raw column and add the formatted one beside it; never overwrite the source
- Test on one row first: name splitting and JSON parsing both break on messy data
- Pair the row-changing actions — Delete Row, Send To Another Table — with a tested condition before turning on Auto-run
- Name output columns after what they hold, so downstream dynamic variables read clearly