2-Hour AI Procurement System: Replace 8 Excel Sheets with Automated Data Management

Many procurement departments manage daily operations with multiple Excel sheets but lack unified data relationships. The author observed a department using eight separate Excel files: Purchase Request, Supplier Archive, Price Comparison, Purchase Order, Receipt Record, Payment Plan, Performance Evaluation, and Complaint Feedback. These sheets suffered from inconsistent supplier names, duplicate data entry, and manual consolidation whenever management needed answers.

To solve this, the author built a complete procurement system in two hours using JianDaoYun, a low-code platform. The implementation followed four main steps: designing structured forms, automating approval workflows, adding AI-assisted analysis, and creating management dashboards.

Building Procurement Forms

Instead of simply moving Excel online, the author reorganized fields and data relationships into five core forms:

Purchase Request Form : Captures requesting department, requester, item, specification, quantity, estimated unit price, expected delivery date, budget allocation, and purpose. Quantity and unit price use numeric fields; estimated total amount is auto-calculated via formula.

Supplier Archive Form : Records supplier name, unified social credit code, contact, cooperation category, qualification documents, access status, and risk level. The key goal is to unify supplier information — each supplier gets a single archive to prevent naming inconsistencies across different users.

Purchase Record Form : Includes purchase number, linked request, linked supplier, category, quantity, unit price, total amount, expected delivery date, and purchaser. It links directly to the request and supplier archive, eliminating re-entry of existing information.

Performance Record Form : Logs actual delivery date, actual quantity, delay flag, quality pass flag, exception description, and handling result.

Complaint Feedback Form : Tracks complaint type, responsible supplier, linked purchase order, handler, handling status, and rectification result.

Original receipt records, payment plans, and performance evaluations were integrated into the relevant forms and process steps. Using JianDaoYun's linked forms, purchase records, performance records, and complaints all connect to the same supplier archive, turning each supplier into a complete business profile with accumulated purchase amounts, delivery performance, and complaint history.

Connecting Procurement Workflow

After forms were ready, the author moved the approval process — previously handled via WeChat, email, and offline signatures — into the system. The workflow follows the actual procurement sequence: Purchase Request → Department Approval → Purchase Execution → Receipt Confirmation → Performance Evaluation → Complaint Handling. Configuration involved three steps:

Set approval conditions : Requests go to the department head first. If the estimated amount is under 5,000, approval routes directly to procurement; if over 5,000, a general manager approval node is added. Rejections return the request to the applicant for revision, preventing unapproved purchases from entering execution.

Set data linking : After request approval, procurement creates a purchase record linked to the original request. Purchase records link to supplier archives; receipt records link to purchase orders; performance evaluations and complaints link to both the supplier and the corresponding purchase record. This enables full traceability from request to delivery.

Set node reminders : Pending approvals remind the responsible approver; upcoming delivery dates remind procurement to follow up; quantity mismatches, delays, or quality anomalies trigger notifications to relevant managers. Every node now has an owner, a status, and a recorded outcome, reducing manual follow-ups and missed information.

Integrating AI Capabilities

With data fragmentation solved, the author added AI assistance to help procurement staff not only query data but also organize and analyze it faster. Three scenarios were implemented:

Auto-extract procurement requirements : Employees often write lengthy descriptions in request notes. AI extracts category, specification, quantity, and urgency, converting unstructured text into structured fields for easier processing.

Assist supplier recommendation : Procurement decisions should consider historical performance, not just price. The system combines supplier archives, historical purchase records, performance evaluations, and complaints. For example, Supplier A may offer a lower price but have recent delays, while Supplier B is slightly higher but delivers on time with fewer complaints. AI summarizes this comparison, eliminating the need to manually review multiple historical spreadsheets.

Identify supplier risk : If a supplier shows increased complaints or consecutive delays, the system generates risk alerts based on historical data. It can also incorporate external public information (business registration, legal records) to assess operational risk. The author emphasizes that AI provides organized information and risk references; final supplier selection, access, and procurement decisions still require human review.

Building Procurement Dashboard

The final step organized accumulated data into metrics managers actually need. Rather than complex charts, the dashboard focuses on high-frequency questions with three core modules:

Procurement Overview : Monthly request count, monthly purchase amount, pending approval amount, and purchase amounts by department. Managers can quickly grasp current scale and fund allocation.

Supplier Analysis : Supplier purchase amount ranking, cooperation count, delayed orders, performance evaluations, and complaint counts. Procurement leads can identify high-value, unstable, or problematic suppliers.

Abnormal Orders : Centralized view of pending orders, delayed orders, quantity anomalies, quality issues, and pending rectification complaints. Procurement leads prioritize handling without reviewing every order manually.

Using JianDaoYun's data management and dashboard functions, data from requests, orders, and supplier performance is aggregated in real time, eliminating month-end manual reporting.

Results After Deployment

Once the system was running, three major changes emerged:

Reduced duplicate entry : Previously, request, order, and receipt data lived in separate Excel files, requiring repeated entry. Now form linking carries request data into purchase records, and receipt/performance data link to orders, cutting duplicate entry and manual verification.

Improved data statistics efficiency : Previously, management queries on amounts, supplier rankings, or delays forced procurement to merge multiple Excel files. Now data accumulates at the moment of business occurrence; metrics are available anytime, not just at month-end.

Data-backed supplier management : Supplier evaluation used to rely on personal impressions. Now purchase amounts, delivery performance, delay records, and complaints provide concrete evidence. Future inquiries, comparisons, and supplier access decisions have complete historical reference.

The base version took two hours to build. Additional modules — contracts, invoices, budget control — can be added as needed. The author concludes that procurement digitalization does not require a large project; start by consolidating scattered tables, linking key processes, and letting data accumulate automatically to reduce repetitive work and gain decision support.