A CSV, a list or a CRM segment in; every row resolved to a domain, filled with firmographics, technographics, funding, news and people through a waterfall of licensed sources, and written back with the source and freshness on each field. The infrastructure under enrichment, run by autonomous AI agents.
Domain resolution, firmographics, technographics, funding, news, relationships, people and verification behind one upload.
Industry, headcount, revenue range, headquarters, stack, funding, news and the people who matter, on one row.
Each field is filled by the first licensed source that has it and checked before the next source is asked, so a row costs what it needs.
An email or phone lookup that finds nothing bills nothing; only found, verified contact rows count.
A conference export, a partner list or a webinar sheet arrives as names and free-text titles. Before anything can be enriched somebody has to turn "Northwind" into northwind.com, and the sheet has eight hundred of them.
The enrichment tool fills industry and headcount on the companies everyone already knows and leaves the long tail blank. Buying a second tool for the gaps means a second export, a second column mapping and two invoices.
Headcount fields still say what they said at import. The company that doubled, raised a round and swapped its stack looks exactly like it did the day the record was created, so scoring and routing run on history.
Company names, ambiguous brands and email domains resolved to a canonical website and legal entity before enrichment starts, with a confidence score and the alternates kept where a name matched more than one company.
Industry, headcount, revenue range, headquarters, founding year and the full technology stack, each field filled by the first licensed source that has it and cross-checked against the next when sources disagree.
The latest round with date, amount and investors, dated news such as launches, expansions and partnerships, and the parent, subsidiaries and partners around each company, all as fields on the same row.
Contacts on the sheet enriched with current title, seniority, department and company, verified work email and direct phone where a licensed source has one, and the buying committee added where the row only had a company.
Drop a CSV, choose a saved list or pick a CRM segment. Column mapping is inferred, so a sheet with "Company" and "Email" is enough to start.
Every row gets a canonical domain and entity. Ambiguous names are resolved with the other columns as context and flagged when confidence is low.
Firmographics, technographics, funding, news and relationships filled field by field through the waterfall, with the source and retrieval date beside each value.
Existing contacts refreshed and verified; the missing roles at each company found and added, with emails verified before they count.
Disagreements between sources reconciled, changed fields highlighted against the previous values, unresolved rows held aside with the reason.
A diff of every field that would change in the CRM, staged as an approval batch; or a new CSV, sheet or list with the original columns intact.
Six hundred attendee names and companies from an event export, resolved to domains, enriched with size, industry and stack, contacts verified, and the rows outside your ICP marked so the follow-up starts with the right hundred.
Every account owned by one team re-enriched on a schedule: headcount, funding and stack refreshed, changes highlighted, and only the fields that actually moved staged for write-back.
A named-account list enriched with the technologies each company runs, when they were adopted and what changed in the last 90 days, so the pitch references the tool the buyer just replaced.
Free-email signups resolved to the company behind them where a licensed source has the link, enriched with firmographics and scored, so the sales-worthy signups surface within the hour.
| By hand | With AstroFabric | |
|---|---|---|
| Domains | Looked up by hand or guessed from the name | Resolved with a confidence score and alternates kept |
| Coverage | One vendor, half the rows blank | A waterfall of licensed sources, field by field |
| Provenance | A value with no idea where it came from | Source and retrieval date on every field |
| Freshness | Enriched once at import | Refreshed on a schedule, changes highlighted |
| People | A separate tool and a separate export | Contacts verified and the missing roles added on the same run |
| Write-back | CSV upload, column mapping, overwrites | A field-level diff, approved once, upserted |
A sheet of company names resolved to domains, enriched with firmographics, stack and funding, the buying committee added and verified, returned as a list and a CSV.
One CRM segment re-enriched every month, headcount, funding and stack changes highlighted, the field diff staged for approval in HubSpot or Attio.
New signups resolved to their company, enriched and scored against the ICP within the hour, the qualified ones pushed to the CRM with the reasons.
Three of a hundred. The full playbook library ships in every workspace, and anything you can describe becomes a mission.
Resolving, enriching and verifying run inside the workspace credit ceiling with no gate. A write-back into the CRM is staged as a field-level diff for a person to approve.
Every field that would change shows its old value, new value and source before approval, and the previous values stay in the run history afterward.
The source capability and the retrieval date ride with each value, so a rep can tell a fresh headcount from a stale one at a glance.
Company and people data comes from licensed providers. The agents never scrape member networks or send messages in your name.
The layer that takes any file, list or CRM segment and returns it complete: names resolved to domains, firmographics, technographics, funding, news and people filled through a waterfall of licensed sources, every field carrying its source and date, written back after approval. Autonomous agents run it end to end.
For each field, the agents ask the licensed source most likely to have it, check the answer, and only ask the next source when the first returned nothing or disagreed with what is already known. Fewer calls per row, better coverage on the long tail, and a source recorded on every value.
Domain and legal entity, industry, headcount and growth, revenue range, headquarters and locations, founding year, funding stage and rounds, technologies in use and recent changes, news, parent and subsidiary relationships, and for people the current title, seniority, department, verified email and direct phone.
Yes. Names are resolved to canonical domains first, using any other columns as context, with a confidence score per row. Ambiguous matches keep their alternates and are held for review, so a common name never quietly enriches the wrong company.
The agents build a field-level diff against the current records in HubSpot, Pipedrive, Attio or Zoho CRM, showing old value, new value and source. A person approves the batch once and only changed fields are upserted; the previous values remain in the run history.
Company enrichment is a few credits per row, a verified contact is a few credits, and resolving, checking and exporting are free. Every rate is shown before the run starts, misses on contact finding cost nothing, and plans start at $49 per month.
Yes. list_enrich, company_lookup, tech_stack and person_enrich are tools on the REST API, the MCP server and the CLI, so a script or an agent you already run can enrich a batch and read the same rows, sources and dates the console shows.
Upload the sheet. Every row comes back resolved, enriched and sourced, and the CRM write-back waits for one approval.