Which Vendors Are Really Worth Choosing? Here's How AI Vendor Scoring Helps Procurement

How Can Vendors Be Assessed More Objectively?

AI Automationcan help procurement transform vendor data into scores based on multiple parameters, rather than just looking at quotation prices. This approach is useful when a company has multiple suppliers with varying performance.

Pricecan be given a weight of 40%.
Quality of product or servicecan be given a weight of 25%.
Lead timeby 15%.
SLA and service responseby 10%.
Warranty or after-sales serviceby 10%.
Vendors with low scores can be directed toreview or re-evaluation.

For example, if there are 20 active vendors, Procurement can create categories.≥90 = Preferred, 75–89 = Approved, 60–74 = Monitor, And <60 = Review. This figure is an example of a threshold that can be adjusted according to company policy.

Why Doesn't Low Price Always Mean the Best Vendor?

In the field, the vendor with the lowest quotation does not necessarily produce the lowest costs in the long run.

For example, Vendor A offers a price of Rp100 million with a lead time of 3 days. Vendor B offers Rp92 million, but their average delivery time is 10 days. If the delay causes operational delays, the Rp8 million difference may no longer be a profit.

With AI Automation, procurement can combine quotation data with transaction history and vendor performance.

1. Determine the Parameters that Can Be Measured

Scoring will be more useful if the criteria have clear data.

For example:

Price → Total Cost

Performance → Delivery & Quality

SLA → Response & Resolution Time

Commercial → Payment Terms

Service → Warranty & Support

With AI Automation, the data can be collected from ERP, procurement systems, spreadsheets, or internal databases.

2. Calculate Performance Based on Actual Data

Don't just use subjective judgment.

For example, a vendor has20 shipping transactions. A total of 18 were completed according to SLA and 2 were late.

Compliance rate:

18 ÷ 20 × 100 = 90%

If the company's target is a minimum of 95%, the vendor needs to be in the categoryMonitor, even though the product prices are competitive.

With AI Automation, calculations like this can be done periodically from actual transaction data.

3. Include SLA in Score

SLA must have clear indicators.

For example:

Maximum response time:4 hours
Resolution target: 2 working days
Delivery target: 3 working days

If the average vendor responds within 8 hours, there is a gap in the SLA.

With AI Automation, the system can read the data and provide a flag when performance passes the threshold.

4. Combine All Parameters

Simple example:

Parameter	Weight	Score Vendor A
Price	40%	90
Quality	25%	85
Lead Time	15%	80
SLA	10%	95
Warranty	10%	90

Final score:

(90×40%) + (85×25%) + (80×15%) + (95×10%) + (90×10%) = 88,25

The vendor is in the categoryApprovedif the company threshold sets 75–89 as a vendor that is still suitable for use.

With AI Automation, the formula can be run automatically every time the vendor data is updated.

What needs to be prepared for Procurement?

Define 4–6 main parameters.
Assign weight based on business priority.
Use actual transaction data.
Set threshold score.
Separate new vendors and existing vendors.
Create Preferred, Approved, Monitor, and Review categories.
Determine the evaluation period, for example everymonth or quarter.
Save the reason for the score change.
Create a notification if the SLA or score drops past the threshold.
Continue to conduct human reviews for strategic decisions.
Use legal software for companies and officially licensed software.

With this model, vendor scoring is more than just a report. The results can be used to determine which vendors to retain, negotiate with, or replace.

FAQ

1. What is vendor scoring?

Vendor scoring is a method of assigning values to suppliers based on parameters such as price, quality, performance, and SLA.

2. Should price be the biggest weight?

Not always. If quality and SLAs are more critical, the weighting can be increased.

3. How often do vendors need to be evaluated?

For active vendors, monthly or quarterly evaluations are quite common. High-risk vendors may be monitored more frequently.

4. Can SLA be calculated automatically?

Can. AI Automationcan calculate response time, delivery compliance, and other SLA indicators if the data is available.

5. What if the vendor performance data is incomplete?

Vendors can be assigned statusNeed Reviewor temporary scores. Don't create scores as if the data were complete.

6. Can new vendors be scored immediately?

Yes, but the criteria should be different. New vendors can be assessed based on quotations, documentation, certifications, capacity, and initial verification results.