Use cases

Clearbit research patterns for distinct go-to-market teams

Use the same evidence discipline differently: a founder needs focus, RevOps needs governance, an SDR manager needs reviewable daily inputs, and a GTM engineer needs explicit data behavior.

Business question

Which observable characteristics distinguish accounts the team can serve now from aspirational accounts that consume research time?

Research pattern

Define firmographic and technographic criteria, exclusions, regional boundaries, and a recency window. Ask the agent for a small candidate set with a reason beside every match. Review false positives before expanding.

Control

Do not treat funding, hiring, or website visits as proof of purchase intent. Record them as signals with dates and competing explanations.

Business question

Which missing company and role fields are necessary for routing, segmentation, and reporting?

Research pattern

Map each destination field to an accepted source, freshness rule, fallback order, and “unknown” state. Use waterfall enrichment only where the added source improves a documented decision.

Control

Keep raw source context and transformation logic available for audit. Deduplicate on resolved company identity, not a normalized name alone.

Business question

Can a representative understand why this account and role are relevant without opening ten browser tabs?

Research pattern

Produce a compact brief containing identity, role rationale, dated trigger context, approved claims, and unresolved questions. Draft outreach only after the account passes review.

Control

Keep human-in-the-loop approval. Check SPF, DKIM, DMARC, suppression lists, opt-out language, and applicable regional rules before sending.

Business question

Can research run where the team already works while permissions, secrets, outputs, and failure states remain understandable?

Research pattern

Install in a test environment, connect the minimum required data source, and run a known cohort. Observe package behavior, network calls, output schema, retries, and deletion steps.

Control

Keep API keys in the runtime secret store. Set cost limits, rate limits, field-level acceptance checks, and a rollback path before production use.

Run one use case against a known cohort

Copy the command, install in a test environment, then compare the result with records your team has already reviewed.

npx -y @okki-global/okki-go-taroball