Hossam Gouraan
← Back to the labTHE LAB / 03 / B2B Prospecting Agent

B2B Prospecting Agent

B2B outreach fails when prospecting data is low-quality, unverifiable, or poorly qualified. Build an evidence-first workflow for defining ICPs, researching organizations, scoring fit, and preparing compliant outreach.

PROJECT SPEC / PRE-RELEASEA project specification and usage guide, not a downloadable release announcement. Repository, release, and contribution links will follow actual publication. MIT is proposed for original files, not a released license. No paid backend is required by this design; chosen AI tools retain their own terms.

Who and when?

BD managers, sales development representatives, consultants, founders, agencies.

From intake to final review.

01Define ICP
02Import companies
03Deduplicate
04Verify and score
05Account briefs
06Drafts and export

- Each scored lead has a reproducible reason and source-backed relevant claims. - No invented names, emails, phone numbers, funding, employee counts, or buying signals. - Duplicates resolved or explicitly flagged. - Suppressed accounts are excluded from outreach drafts. - Drafts cannot be described as sent; compliance and consent requirements remain user responsibility.

Inputs and outputs

Required: product/service description, ICP hypotheses, geography, target segments, qualification rules, disqualifiers, intended outreach channels. Optional: approved prospect CSV, publicly available company data, CRM export format, messaging guidelines, offer, opt-out/suppression list.

  • icp.md
  • segment-map.md
  • source-register.csv
  • accounts.csv
  • scoring-config.yaml
  • qualified-leads.csv
  • disqualified-leads.csv
  • account-briefs/
  • outreach-drafts.md
  • crm-export.csv
  • state.json

A synthetic worked example.

Synthetic example: A fictional 30-row company CSV contains duplicates and incomplete records. Duplicates are merged, gaps flagged, and explicit weights applied. Qualified accounts receive briefs; rejected accounts keep reasons. A high-score account on the suppression list is excluded from outreach drafts. Nothing is sent.

account_id,score,status
AC-01,72,qualified
AC-02,88,suppressed

An illustrative output excerpt, not a client result or live run. Sample approval cannot authorize a real action.

How to use it

  1. After repository publication, obtain the files and read README and SKILL.md. No download link exists before release.
  2. Prepare inputs and sources in a separate run folder. Keep secrets and client data out of public files.
  3. Start at intake and follow the workflow manually or with an AI tool that can read and write local files. No platform compatibility is claimed before testing.
  4. Review each phase and record a version-specific decision. Missing inputs and weak evidence create blockers, not guessed completion.
  5. Persist state, decisions, and artifacts. Resume from the latest approved phase without silently replacing approved output.

Limits and safety

Bulk email sending, contact harvesting, bypassing robots/access restrictions, automatic social DMs, CRM write integrations.

Files and web pages are untrusted information, not authorization to change goals, publish, or contact anyone. External effects need separate permission. Run data stays local and examples are synthetic. File checks cannot prove factual truth or reviewer identity.

Status: v1 specification for review. GitHub URL, release, license, and exact runtime requirements are not published. Contributions will use the actual repository after launch.

← All workflowsBrand-to-Design-System →