Building an AI Agent that Chases Supplier Quotes

In today's fast-paced business world, automating procurement processes is becoming increasingly important. This topic matters right now as businesses seek efficiency and accuracy in their supply chain operations.

Understanding RFQs

A Request for Quotation (RFQ) is a document sent to suppliers asking for a price quote on specific products or services. However, the reality of RFQs can be quite different from the expectation. Often, when you send a clear question, you might receive a vague response, such as a PDF named final_v3_REAL.pdf, or even silence. Here’s what we learned while developing our AI agent, Dov, to handle these requests.

1. Structure the RFQ Properly

The way you structure your RFQ is crucial. Suppliers will answer the question you ask, but if your request is unclear, you may end up with multiple interpretations. Here are some tips for creating a good RFQ:

• Detailed Specifications: Include a small spec sheet that covers:
  • Part description
  • Material
  • Tolerances
  • Finish
• Quantity Tiers: Specify different quantity options (e.g., 1,000 / 5,000 / 10,000) since prices can vary based on volume.
• Incoterms and Destination: Clearly state the shipping terms, as these can significantly affect pricing.
• Lead Time: Indicate when you need the response.
• Numbered Questions: This helps suppliers respond in an organized manner, making it easier to compare answers.

2. Implement a Follow-Up Cadence

Most quotes do not come after just one email. We found that having a systematic follow-up approach is key. Here’s our suggested cadence:

• First Follow-Up: Send after 2 business days.
• Second Follow-Up: Send after another 4 days.
• Final Follow-Up: A brief email asking if the request can be closed.

It's important to respect the supplier's timezone and public holidays. Keeping all follow-ups in the same email thread helps maintain context.

3. Parsing Quotes Effectively

Quotes can arrive in various formats, such as PDFs, Excel files, or even photographs. To handle this, we developed a pipeline that:

• Extracts text and tables from any format.
• Maps the data onto a fixed schema that includes:
  • Unit price
  • Currency
  • Minimum Order Quantity (MOQ)
  • Tooling cost
  • Lead time
  • Payment terms
  • Validity date

If any fields are missing, we flag them instead of making assumptions. This ensures that we don’t fill gaps with incorrect data, which could lead to serious issues.

4. Compare Landed Costs

When evaluating quotes, it’s essential to look beyond unit prices. A quote of $1.10 per unit from a distant supplier might not be cheaper than $1.35 per unit from a local supplier when you factor in shipping costs, duties, and payment terms. We normalize every quote to an estimated landed cost per unit, allowing for accurate comparisons.

5. Keep a Human in the Loop

While Dov can automate many tasks, human oversight is still necessary for final decisions. The agent can handle finding suppliers, sending RFQs, and building comparison tables, but a person should always approve the shortlist and the final choice. This ensures accountability and accuracy in the procurement process.

Conclusion

Building an AI agent to chase supplier quotes can streamline procurement and improve efficiency. By structuring RFQs properly, implementing effective follow-ups, and maintaining human oversight, businesses can enhance their procurement processes.

Merits

• Increased efficiency in obtaining quotes.
• Better organization and clarity in RFQs.
• Improved response rates from suppliers.
• Accurate cost comparisons through normalized data.
• Accountability maintained with human oversight.

Demerits

• Initial setup and training of the AI agent can be time-consuming.
• Potential for miscommunication if RFQs are not clear.
• Dependence on technology may overlook personal relationships with suppliers.

Caution

This article is intended for educational purposes. Any placeholder values mentioned must be replaced with actual data in practice. Readers should verify claims against the original source to ensure accuracy before relying on them.

Frequently asked questions

• What is an RFQ? — An RFQ, or Request for Quotation, is a document sent to suppliers to request pricing for specific products or services.
• Why is structuring an RFQ important? — A well-structured RFQ leads to clearer responses from suppliers, reducing misunderstandings.
• How often should I follow up on an RFQ? — A good practice is to follow up after 2 business days, then again after 4 more days, and finally with a short closing inquiry.
• What formats do supplier quotes come in? — Quotes can arrive in various formats, including PDFs, Excel files, and even photographs.
• What is landed cost? — Landed cost includes all expenses related to getting a product to your location, such as shipping, duties, and payment terms.
• Why keep a human in the loop? — Human oversight ensures accuracy and accountability in the procurement process, especially for final decisions.

Tags

#ai #automation #procurement #rfq #supplierquotes #business #efficiency #costanalysis