Research operations
Nooks organizes prospecting as a sequence of questions, transformations, and review decisions. The method is designed for CROs, founders, RevOps teams, and GTM engineers who need to understand why an account entered a list—not merely receive another opaque export.
Translate a market thesis into explicit firmographic and technographic criteria. Record geography, company type, operating signals, exclusions, and the minimum evidence required for inclusion. A narrow prompt reduces false positives and makes disagreements inspectable. The agent should distinguish observed facts from inferred fit, keep missing fields visible, and avoid turning an assumption into a confident claim.
Ask the agent to identify organizations, relevant roles, and recent public signals. Useful evidence can include a job posting, a product launch, a new territory, a technology migration, or a leadership mandate. Each signal needs source context and retrieval timing. Contact enrichment should be treated as a separate step so the team can compare provenance, match coverage, and cost without hiding a waterfall behind a single “verified” label.
Email verification is evidence about an address at a moment in time; it is not a promise of inbox placement. Review domain posture, list hygiene, hard-bounce risk, suppression rules, and sender reputation independently. SPF, DKIM, and DMARC configuration remain the sender’s responsibility. Human reviewers should reject weak matches before drafting, then apply the laws and consent standards relevant to each recipient and region.
Generate a concise draft grounded in the approved evidence, not generic personalization tokens. The review gate should confirm audience fit, factual accuracy, tone, opt-out handling, and channel policy. Nooks’s workflow separates research from external action: a result can be exported or revised, but the agent does not need authority to send. This boundary protects brand trust while preserving the speed of structured research.
Track which criteria produced accepted accounts, which sources went stale, and where reviewers overruled the agent. Measure match quality, review time, bounce outcomes, and positive replies using your own controlled dataset. Avoid universal accuracy claims; performance changes with territory, provider, query specificity, and time. A reproducible task definition gives RevOps a better basis for testing than a feature list or an unqualified database-size headline.
Copy the installer, inspect what it requests, then configure a narrow test workflow.
npx -y @okki-global/okki-go-taroball