Skip to content

account_research

account_research is the one-call account brief. It reads the tenant’s account_research configuration, derives the in-scope use cases, then pulls 10-K extracts, the employee technology footprint, intent signals, and AI-scored prospects in parallel, and ships everything plus a rendering contract in a single JSON envelope.

It replaces the older bdr_account_research tool.

  • “Generate a report on Acme.” / “Research this account.” / “Build a BDR brief for acme.com.”
  • Any request that pairs a company domain with use cases, intent signals, or 10-K context.
ParameterTypeRequiredDescription
company_websitestringYesTarget company website, e.g. "capitalone.com".
tenant_idstringYesTenant whose configuration drives keyword choice and prospecting enablement.
company_linkedin_urlstringNoRequired for the LinkedIn footprint section and for AI prospecting.
include_prospectsbooleanNoSkip the prospecting run when false. Default true.
signals_sincestringNoISO date floor for signals, e.g. "2025-01-01".
signals_limitintegerNoRow cap for the signals pull. Default 100.
filings_limitintegerNoFilings to consider. Default 1 (most recent).
footprint_limitintegerNoRow cap for the footprint pull. Default 30.

One envelope containing:

  • The assembled research sections — tenant fit / use cases, 10-K extracts, employee technology footprint, intent signals, AI-scored prospects.
  • render_spec — the rendering contract for the customer-facing report.
  • datasets — a flat map of dataset ids (filings_10k, linkedin_footprint, intent_signals, prospects) so you can slice any section further with query_datasets without re-running the tool.

Prospecting is asynchronous (one to three minutes). If it hasn’t finished within the per-call budget, the response carries status="still_running" — call again with the same arguments to resume.

  • One call does the whole fan-out; don’t hand-orchestrate the individual pulls.
  • Need a 10-K section it didn’t surface by default? Follow up with query_company_filings.
  • Need a deeper competitor footprint cut? Follow up with ask_onfire using a technology insight_filter.