Guided workflow
Move from a falsifiable ICP to sourced account research, controlled enrichment, and human-approved action. Each state remains visible so speed never hides provenance or responsibility.
Translate the market thesis into firmographic and technographic criteria, territory, exclusions, evidence thresholds, and intended use. Observed facts stay separate from inferred fit. Missing fields remain visible rather than becoming confident guesses, giving RevOps a stable brief to reproduce and audit.
Research accounts, roles, hiring activity, product changes, and other public intent signals while recording source context and retrieval timing. Treat contact enrichment as a distinct stage. A waterfall may increase field coverage, but it also adds provider cost, conflict resolution, rate limits, and deletion paths.
Keep targeting, contact choice, message drafting, suppression, and sending as separate decisions. A human reviewer checks relevance, factual accuracy, regional rules, provider terms, and approved channels. Verification reduces one uncertainty; it never guarantees inbox placement, replies, consent, or compliance.
Use a labeled set to measure accepted matches, evidence completeness, reviewer overrides, research time, provider spend, and downstream outcomes. Avoid universal accuracy claims because results change with segment, geography, query design, source freshness, runtime, and reviewer standards.
Broad criteria, mixed sources, hidden transformations, uncertain freshness, and sending permissions bundled into the same automation make errors difficult to isolate.
A written brief, sourced result, reviewer decision, and separately authorized action create an evidence trail. Unknown remains a valid outcome when public evidence is insufficient.
Copy the installer, inspect the package, and configure only the sources needed for a labeled test.
npx -y @okki-global/okki-go-taroball