Enrich New Salesforce Accounts with Firmographic Research

This workflow turns a bare-bones new Salesforce account into a fully researched record automatically. The moment a sales rep or inbound form creates an account with little more than a company name and domain, the workflow fires off live web research with Perplexity, hands the raw findings to OpenAI for structured extraction, and patches the verified firmographics straight back onto the Salesforce account — industry, employee headcount band, estimated annual revenue, HQ location, latest funding round, and a two-sentence sales talking-point brief.

No more reps tabbing out to LinkedIn, Crunchbase, and the company website to hand-fill custom fields. Every new account lands enriched, normalized to your picklist values, and ready for routing or scoring.

Reach for this template when:

• Inbound leads and self-serve signups create thin Salesforce accounts that your routing, scoring, or territory rules need firmographics to act on.
• Your SDRs waste the first ten minutes of every account on manual company research before they can personalize outreach.
• You want consistent, picklist-normalized values for Industry, Employee Count, and Revenue instead of whatever a rep typed by hand.
• You need a fresh, source-backed "what does this company do" brief attached to the account before the first call.

Skip it if your accounts are already enriched upstream by a paid data vendor and written through that vendor's native Salesforce sync.

1. 1A new Account record in Salesforce triggers the workflow and passes its name and website domain.
2. 2Perplexity runs live web research on the company — current industry, size, revenue, HQ, recent funding and news — returning cited findings rather than stale cached data.
3. 3OpenAI receives the research text and the account's existing fields, then extracts a strict JSON object: normalized industry, employee band, revenue estimate, HQ city/country, last funding round, and a concise sales brief. It is instructed to leave fields null rather than guess when the research is inconclusive.
4. 4The workflow updates the Salesforce account, writing each extracted value to its mapped field and the brief to a long-text "Account Research" field, skipping any field the model returned as null so it never overwrites good data with blanks.

Because it runs on Agent Hive's Sim engine, the research and extraction steps execute as real, observable runs — not a brittle screen-scrape — so you can inspect every input and output when a value looks off.