National Project Opportunity Recommendation Assistant

The Problem It Solves

Bidding data is scattered across multiple platforms and announcements, making manual monitoring of nationwide project opportunities inefficient for enterprises. Key challenges include:
- Massive data volume, with over 300,000 updates daily, hindering real-time tracking.
- Complex query conditions, involving dimensions like region, industry, budget, and procurement units.
- Time-consuming analysis to filter high-value opportunities (e.g., near-construction projects, contract renewals) from raw data.

Core Capabilities and Workflow

This skill converts vague business needs into actionable searches via natural language interfaces and structured processing. The core workflow is:
1. Intent Recognition and Parameter Extraction: Parses user input to extract keywords, regions, industries, and time ranges. For example, it automatically converts “find smart park tenders in East China over the last 90 days” into API query parameters.
2. Data Retrieval and Filtering: Uses whitelisted interfaces (e.g., SearchProjectForAI, searchNZJProjectApi) to fetch data, applies default rules (e.g., 30-day timeframe if unspecified), and filters irrelevant content (like job postings or news).
3. Analysis and Output: Generates a “30-second actionable opportunity brief” including:
- One-sentence conclusion: Quick judgment on project priority.
- Key information: Lists project name, budget, procurement unit, source link, etc., with traceable fields.
- Opportunity assessment: Analyzes from angles like demand match, competition intensity, and procurement stage.
- Follow-up actions: Provides executable steps, such as contacting procurement units or preparing materials.
- Further exploration: Suggests next steps, like querying competitor history or exporting project lists.

Applicability and Limitations

This skill focuses on bidding data analysis with clear boundaries:
- Data dependency: Relies entirely on the Shipboy Technology API (e.g., endpoint /outer-gateway/bid/{interface}), requiring a valid BBIAO_API_KEY environment variable. Non-configured setups will not function.
- Interface whitelist: Only calls SearchProjectForAI, searchNZJProjectApi, getNZJProjectDetail, searchProjectContactApi, and getCollectUrl; no other interfaces are extended.
- Output nature: Provides data-driven analysis and recommendations but does not guarantee winning bids or circumvent rules; all figures and dates must come from API responses without fabrication.
- Handling quota issues: On API error code 0100590006 (insufficient balance), it prompts the user to recharge and halts current calls without automatic retries.