How to build a Tender Intelligence Agent
This agent automates the extraction, analysis, and reporting of all critical information from tender documents.
Challenge
Tender managers waste hours manually extracting, summarizing, and reporting on key details from complex logistics tender documents.
Industry
Logistics
Department
Content Creation
Legal
Integrations
OpenAI
Google Docs
Workflow Overview
1. File Upload & Preprocessing
Files Node
Purpose: Users upload tender documents (PDFs, scans, etc.).
What happens:The node processes files, including OCR for scanned documents, and extracts the text for downstream analysis.
2. Document Formatting
Document Markdown Extractor (Template Node)
Purpose: Converts the extracted text from the uploaded files into markdown format.
Why:Markdown formatting ensures the text is clean and structured for the AI models to analyze.
3. Section-by-Section AI Extraction
A series of specialized AI (LLM) nodes each focus on extracting a specific section from the tender document:
Node Name | What it Extracts |
|---|---|
Name of Company | The company’s name from the document. |
Pickup & Delivery Locations | All pickup and delivery locations mentioned. |
Special Terms & Conditions | Any special terms and conditions, summarized. |
Pricing & Payment Terms | All pricing and payment terms, clearly listed. |
Key Dates & Deadlines | All important dates and deadlines. |
Risks & Liabilities | Main risks and liabilities described in the tender. |
Contact Points | All contact points and responsible parties. |
How it works:
Each LLM node receives the markdown-formatted document and a specific prompt tailored to extract its section.
4. Section Outputs
Output Nodes
Each LLM node’s result is sent to a dedicated output node, so you can view each extracted section individually.
5. Executive Report Generation
Report (LLM Node)
Purpose: Compiles all the extracted sections into a single, well-structured executive report.
How:The node takes the outputs from all previous LLM nodes and merges them into a professional report, using clear section headers.
6. Google Doc Creation
Create Google Doc (Action Node)
Purpose: Automatically creates a new Google Doc containing the full executive report.
How:The file name is set to “Tender Report: [Company Name]”.
The content is the full report, including citations if available.
7. (Optional) Knowledge Base Node
Knowledge Base Node
Purpose: Present in the flow, but not currently connected.
Potential use: Could be used to search or cross-reference information from a knowledge base in future enhancements.
Key Points
Automated: Upload a document, and the workflow extracts, summarizes, and compiles all key information.
Modular: Each section is handled by a dedicated AI node, making it easy to review or modify extraction logic.
Output: You get both individual section outputs and a polished executive report, automatically saved to Google Docs.
