The problem this solves
Every team has the spreadsheet: names typed from a badge scan, a partner's export, a list bought two years ago. The company column is a mix of brand names and legal names, half the rows have no domain, and the columns that would make the rows useful - size, industry, stack, funding, a working email - are empty. Enriching it means pasting into three tools in sequence and paying each of them for the rows the previous one already filled.
The waterfall is the fix for both the gaps and the cost. Each field is tried against one source, and the next source is consulted only where the first came back empty, so coverage climbs field by field and spend stays proportional to what was actually filled. The result is a spreadsheet where every column is as full as the market allows and a report that says exactly how full that is.
How the mission runs
- Read the CSV and resolve domains. The Enrichment Agent reads the attached file, detects which columns hold company names, contact names and any partial domains, and resolves each company to an official domain. Ambiguous names are resolved with the other columns as context, and unresolved rows are marked rather than guessed.
- Fill firmographics through the waterfall. Headcount, industry and HQ country are filled from firmographic sources one after another until each field has a value, and the row records which source answered. A row that no source can fill stays empty with a note, which is more honest than a placeholder.
- Add the stack and the latest round. Technographics supply the detected technologies per company, and funding data adds the latest round with its date and amount. Companies with no funding history are marked as bootstrapped or unknown rather than left blank without explanation.
- Find and verify the contact email. For each named contact, the email finder searches for a work address at the resolved domain and verification checks it. The status - deliverable, catch-all, risky, invalid - lands in its own column beside the address.
- Return the file and the coverage report. The enriched CSV is returned with the original columns preserved and the new ones appended, and a coverage report shows the fill rate per column before and after, so you can see what the waterfall achieved and what the data simply does not offer.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
The original CSV with every row carrying a resolved domain, headcount, industry, HQ country, detected technologies, latest funding round and a verified work email for the named contact, plus a per-column coverage report stating how many rows were filled for each field and which rows could not be resolved. Nothing in the original file is overwritten; new values sit in new columns.
Make it yours
- Drop the funding column and add direct phone numbers when the list is for calling.
- Point the mission at a Google Sheet instead of a CSV and the enriched columns are written into a new tab.
- Add intent topics per company to see who on the list is researching your category right now.
- Save the result as list <name> so it can be refreshed, scored and pushed to a CRM later.
Frequently asked questions
How does the waterfall decide the order of sources?
By historical fill rate and cost for each field type, so the most likely and cheapest source is tried first. You can see which source filled each cell in the report.
What happens with rows where the company cannot be identified?
They are returned unchanged with a note explaining the ambiguity, and you can supply a domain for them and re-run only those rows.
Is the original data changed?
No. Original columns are preserved exactly; enrichment lands in new columns, and a normalized company name is offered as its own column rather than replacing yours.