Join the Community

Optimizing Clay AI: Matching Jobs with Customer Records

·

·

Hey Clay community! Looking for the best approach to solve this use case: Setup:

Table 1 ("Jobs"): 150 job postings with standard Clay enrichments (company, domain, etc.)
Table 2 ("Customers"): 30 existing client records (uploaded CSV)
Both tables in same workbook

Goal: Create a column in the Jobs table that uses Clay AI to:

1.
Reference all 30 records from Customers table
2.
Find the 2 most similar companies based on industry, revenue & ICP
3.
Provide reasoning for each match

Question: What's the best way to structure this? I was thinking "Send Table Data" to reference the Customers table in a Clay AI column, but I have been constantly getting an error: Prompt has the error(s): Rows from: Customers > Company is blank Rows from: Customers > Website is blank I checked thoroughly and confirm that there no empty cells in either the Customers > Company column or Customers > Website column. Thanks in advance!!

2 comments

Abdul S.
·
·
Hey Gagan D. this usually happens if there's hidden whitespace in the Customers table. Try trimming the “Company” and “Website” columns first For setup: Use “Send Table Data” in a Clay AI column in your Jobs table Reference the Customers table Ask Clay to find the 2 most similar companies based on industry, revenue, and ICP with reasoning If you’re still stuck, feel free to DM me happy to help!
❤️1
👋1
Asher
·
·
Gagan D. Hey Gagan, I've done something similar before. The best way might be this:
1.
Manually sort your 30 customer references into different industries and revenue range that relate to the company industries you serve. Put it into a Google Doc. It should look something like:
a.
Company Name, Industry, Revenue, and a quick explanation of what the company does
2.
In the 'jobs' table, enrich the company based on industries and revenue range.
3.
Then create one AI formula column , paste the google docs with all the client information you have, then get AI to use all information about the company, and choose two customer references.
The way you are doing it now, might not be the best way to use Clay. Because essentially you are getting Clay to:
Find information about one company, and go to another table to find the information about the 30 clients, then compare.
From an architecture standpoint and AI context window standpoint, you might not want to do that. You are better off giving your explaination on each client before hand, so all that AI needs to do is just to compare.
👋1
❤️1

Optimizing Clay AI: Matching Jobs with Customer Records | Clay