Sales Lead Mining

The Problem of Precision Customer Acquisition from Public Data

For B2B sales teams, a persistent challenge is efficiently and accurately identifying potential customers (sales leads) from vast, unstructured public data sources like procurement announcements and public tenders. Manual screening is time-consuming and prone to missing opportunities. Simple keyword searches often fail to effectively assess a project's true relevance, its procurement urgency, or the completeness of contact information, leading to misguided follow-up priorities.

How the Sales Lead Mining Skill Operates

This skill automates the processing of structured data via API calls, separating raw facts from analytical judgments to produce leads that are sourced, interpretable, and actionable. Its core workflow includes:

• Keyword Generation & Search: The user's product or service name is expanded into search keywords (keyword), required terms (inCludeKW), and exclusion terms (excludeKW). It then calls the POST /outer-gateway/bid/searchProjectApi endpoint to scan for relevant public tenders within a specified timeframe and region, prioritizing listings with clear procurement intent.

• Incremental Data Enrichment: Basic details (client, budget, region) are extracted from search results. For high-scoring projects, the POST /outer-gateway/bid/getZTBStructreDetail endpoint is called to retrieve structured details like project number, deadlines, and contact persons. For confirmed target companies, separate calls are made to the enterprise profile (POST /outer-gateway/bid/companyProfileSummary) and contact endpoints (POST /outer-gateway/bid/companyProfileContacts) to add background and publicly available contact leads. Crucially, contacts from project details and those from enterprise profiles are sourced differently and must be explicitly labeled as such in the output.

• Quantified Scoring & Tiered Output: Each lead is scored on a 100-point scale across six dimensions: demand match, procurement stage, timeliness, budget match, regional match, and contact completeness. Scoring rules can be customized, missing data defaults to 0 points, and no fields are ever fabricated. The final output categorizes leads into "Priority Follow-up", "Monitor", and "Low Match" tiers, with recommended next actions for each.

Boundaries and Important Considerations

The skill's effectiveness is bound by the following:

• Data Dependence: All leads are sourced from public procurement data APIs (authorized via BBIAO_API_KEY). It cannot access non-public procurement opportunities or private networks. Data timeliness and completeness depend on the upstream platform.

• Compliance and Security: Automatic registration, key guessing, or echoing the API key (BBIAO_API_KEY) is strictly prohibited. All API calls must adhere to rate limits and pagination rules (e.g., max 50 items per search page).

• Mandatory Human Verification: The output must include a "Items Requiring Manual Verification" section. Contact information, procurement status, and budget figures derived from public data or analysis may become outdated or inaccurate. The lead details serve as a starting point for research, not as a definitive final decision-making tool.