CRM enrichment

Keep CRM accounts filled and refresh them when they go stale. New records get approved blanks written from LinkedIn; a record whose last successful fill is older than six months comes back.

What it does #

• Fills approved blanks from LinkedIn: name, domain, website, LinkedIn page, employee count, and LinkedIn company ID as a durable matching key.
• Re-enrolls stale records after six months, so firmographics stay current without manual refreshes.
• Never overwrites a populated value. The CRM stays authoritative.
• Stamps freshness only after a real fill. An already-filled row exits before any paid call.
• Separates enrichment from writeback. The tool returns provider data; the play owns the CRM write.

How it works #

flowchart TD
    model["crm_accounts<br/>HubSpot · Salesforce · Attio"]
    filter["Managed segment trigger<br/>has identifier · stale or never filled · one approved blank"]
    enrich["account_enrichment tool<br/>LinkedIn URL route, else domain route"]
    write["enrich_accounts play<br/>fills approved blanks · stamps freshness"]

    model --> filter --> enrich --> write
    write -->|CRM record ID| model

1. Audit. The agent joins live LinkedIn fields to live CRM properties, flags duplicates and transformations, and waits for approval of the field contract.
2. Build disabled. It adapts infra/index.ts, deploys with the play disabled, and shows Cargo links, target population, and estimated credits.
3. Run. After cost approval, enrichment runs and the agent reports fill rates, outcomes, failures, and actual credits.

Architecture #

Tool and play share one workflow contract, so mappings cannot drift. The play’s filter is its segment; there is no standalone segment.

Placeholders (edit before deploy) #

1. CRM connector and record ID (infra/index.ts): the example is HubSpot (hs_object_id); Salesforce uses Id, Attio its record id. The wrong field targets nothing while the run still looks successful.
2. Field mappings (infra/index.ts): every destination must be a live property on the connected CRM.
3. Freshness fields: cargo_last_enriched_at and cargo_enrichment_status must exist on the CRM object.
4. linkedin_company_id: propose it if no equivalent property exists.

Cost #

The audit makes no paid call: it reads schemas and counts eligibility, so the cost preview lands before any approval.

Enrichment is one LinkedIn call per eligible row, priced by route: enrichCompany for a LinkedIn URL, enrichCompanyFromDomain as the fallback. Run cargo-ai connection integration get linkedin for current unit prices. A row with no identifier never calls; an already-filled row exits as skipped_already_filled.

After a successful fill the record leaves the segment for six months, so the daily schedule only pays for new rows and records coming back due.

Done when #

• Audit JSON, Markdown, and chat summary agree on every count
• The plan shows the account_enrichment tool and a disabled enrich_accounts
• Every destination is a live property on the connected CRM
• Rows with no identifier or no blanks exit before a paid call
• The write matches the audited CRM record ID
• LinkedIn and domain route counts are mutually exclusive and reproduce the credit estimate
• The post-run report shows fill rates, outcomes, failures, and actual credits

Composes into #

account-scoring (a filled book is what the scorer can cite), find-stakeholders (the buyers at every filled account), tam-building (the universe these records join).