← Sticker Wall AGENT · DETAIL

AI Bid Qualification, Performance Mining & Risk Scorer - Monthly Revenue 25k

Workflow: Receives a daily list of suppliers to be evaluated uploaded by purchasing enterprises, along with related bidding docume

AGENT

Key Fields

FIELD STAMPS

IndustrySaaS / Enterprise Software

RegionChina

ScaleSME

ChannelOnline

🔧 Workflow

Receives a daily list of suppliers to be evaluated uploaded by purchasing enterprises, along with related bidding documents and qualification files. The AI first uses OCR parsing tools to extract structured information such as qualification certificates, historical delivery projects, and revenue scale for all suppliers. It then automatically crawls litigation and administrative penalty data from the National Enterprise Credit Information Publicity System and China Judgments Online. Combined with the buyer's custom scoring dimensions (such as qualification compliance, fulfillment capability, and risk level), it automatically calculates weights, and finally outputs a red-yellow-green tricolor scorecard with risk annotations and risk explanations for the buyer's decision-making use.

🛠 Setup Requirements

Requires basic prompt engineering skills, familiarity with typesetting and field rules of bidding documents across different industries, using Coze to build document upload and parsing workflows, integrating RAGFlow for long-document precise parsing, and connecting Octopus Collector to acquire public enterprise risk data. In the initial stage, at least 20 bidding document samples from different industries need to be prepared to debug and recognize accuracy. The overall setup cycle is about 10-15 days, with no complex programming skills required.

🧰 Toolchain

🔧 Coze
🔧 RAGFlow
🔧 DeepSeek API
🔧 Octopus Collector

💰 Revenue

① Buyers purchase supplier evaluation reports per piece (main revenue): purchasing enterprises pay 200 RMB for each basic report, and a monthly revenue of 25,000 RMB corresponds to about 125 deliveries (calculated as 25000 ÷ 200). The exact proportion of this path in the total revenue has no public caliber; ② Expedited items or scoring reports customized according to client-specified dimensions (markup per occurrence): 500 RMB each, 150% higher than the basic tier (calculated as 500 ÷ 200 - 1). How many copies are issued per month is unknown to the outside world, and the proportion is also undisclosed; ③ Monthly service agreement signed with procurement platforms (subscription model): single customer monthly fee of 3,000-8,000 RMB. The number of signed companies cannot be verified, but the creator's self-statement states that the monthly revenue exceeded 40,000 RMB at its peak, and the proportion of the total pie is not specified; ④ Opportunity item - licensing scorecard templates and APIs to procurement platforms: peer cases claim that the error and omission rate of bid clearing information is reduced by 90% (case caliber, without independent review), charged via authorization licensing, with both authorization price and proportion having no figures yet.

💸 Cost

Fixed monthly expenses are about 800-1,500 RMB, among which DeepSeek model API call expenses are about 400-800 RMB, Octopus Collector membership fee is about 99 RMB/month, Coze basic version is free, and cloud storage and computing power expenses are about 300-600 RMB/month, with no other extra expenses.

⏱ Time Investment

Investing 3 hours daily to review AI-generated scorecards, correct recognition errors, and meet customer requirements.

🚀 Getting Started

The first step is to collect public bidding documents and supplier qualification materials from 3-5 target industries (such as infrastructure, manufacturing, and e-commerce), build a basic document parsing and information extraction workflow on the Coze platform, complete test scoring for the first batch of 20 suppliers, polish scorecard templates that meet industry demands, and then directionally push them to target customers through bidding forums and purchasing manager communities, first providing 3 free trial reports to build reputation.

🔑 Keys to Success

✅ The OCR parsing accuracy of bidding documents in different industries must remain stable above 95% to ensure no omissions in the extraction of qualification and performance information
✅ Establish a dynamically updated database of supplier litigation, administrative penalties, and dishonesty records to ensure the timeliness of risk data
✅ Scoring dimensions must align with buyers' core pain points, such as setting weight items for qualification fraud, contract performance default, and litigation records
✅ Establish cooperation with at least 2 procurement communities or bidding platforms to obtain continuous customer traffic

⚠️ 风险

⚠️ Bidding document formats vary greatly across industries, and initial debugging of recognition models may require significant time investment
⚠️ If the scoring report omits major risks (such as supplier dishonesty or major litigation), it will face customer compensation and trust crises
⚠️ If sensitive information such as suppliers' business licenses and qualifications is improperly stored, it may involve data compliance risks

📌 Real Cases

📌 A manufacturing enterprise procurement department used this service to screen candidate subcontractors, issuing 20 reports within 3 days, discovering implicit litigation risks in 3 of them, and avoiding potential losses exceeding 500,000 RMB
📌 An infrastructure procurement platform deployed this scoring system to screen 120 suppliers per month, achieving a risk identification accuracy rate of 93% and saving 20,000 RMB/month in due diligence labor costs
📌 A cross-border e-commerce procurement team used this service to filter logistics suppliers, screening out 7 shell companies with no actual fulfillment capabilities, avoiding supply chain interruption losses of about 300,000 RMB

Sources