Errors in ICP Qualification Column: Need Help Troubleshooting

Errors in ICP Qualification Column: Need Help Troubleshooting | Clay

6 comments

· Sorted by Oldest

Clay
APP
·
·
ICP Qualification errors typically occur due to missing or invalid data in the formula inputs. Check if your seniority, job function, and company size fields have complete data. Try these troubleshooting steps: ** Test for empty fields and boundary values in your conditional statements ** Refresh your browser using Cmd/Ctrl + Shift + R to resolve formula evaluation errors * Right-click the column and select "Run All Rows that haven't run or have errors" If issues persist, request human assistance by clicking "Talk to a person."
Sources: Claybooks: Extract, qualify, and enrich LinkedIn event attendees from any event | Clay.com Conditional statements | Documentation | Clay University Run Options
Gabriel Echarte
·
·
Hi Chris, I'm running some tests on this now.
Gabriel Echarte
·
·
I think it has to do with the checkmark icon tripping up the models. I'm changing the output to "success" and "fail" to see if I can get it working reliably.
Gabriel Echarte
·
·
I streamlined the prompt a little using our metaprompter. It's outputting Success or Fail now in a test duplicate column. Do you want to look through this and confirm it's what you wanted? Attaching explanation of issue below!
Gabriel Echarte
·
·
Why This Happened:
1. OpenAI Rate Limits ** Your API key has usage limits (requests per minute/day) ** Claygent makes multiple API calls per LinkedIn analysis ** When limits are hit, requests get queued or fail silently 2. Claygent Timeout Issues ** Each Claygent task has a timeout limit (usually 60-120 seconds) ** Profile scraping takes time, especially for complex profiles ** If analysis takes too long, Claygent stops and returns blank
How to Confirm It's Happening: Check OpenAI Usage: 1. Log into OpenAI dashboard 2. Go to Usage > Activity 3. Look for rate limit errors or high request volume 4. Check if you're hitting monthly/daily limits
How to Fix It:
Option 1: Reduce Batch Size ** Run 10-20 rows at a time instead of hundreds ** Wait 30-60 seconds between batches ** This prevents overwhelming OpenAI's rate limits Option 2: Upgrade OpenAI Tier ** Higher usage tiers get better rate limits ** Check your OpenAI tier status and consider upgrading Option 3: Simplify Prompt Further ** Our streamlined prompt still might be too complex
Hasnain R. (LeadsOps)
·
·
Hey Chris Balestras! It could be due to missing inputs a broken prompt formula issues or failed columns it depends on double-check those and re-run.

Clay
APP
·
·
ICP Qualification errors typically occur due to missing or invalid data in the formula inputs. Check if your seniority, job function, and company size fields have complete data. Try these troubleshooting steps: ** Test for empty fields and boundary values in your conditional statements ** Refresh your browser using Cmd/Ctrl + Shift + R to resolve formula evaluation errors * Right-click the column and select "Run All Rows that haven't run or have errors" If issues persist, request human assistance by clicking "Talk to a person."
Sources: Claybooks: Extract, qualify, and enrich LinkedIn event attendees from any event | Clay.com Conditional statements | Documentation | Clay University Run Options
Gabriel Echarte
·
·
Hi Chris, I'm running some tests on this now.
Gabriel Echarte
·
·
I think it has to do with the checkmark icon tripping up the models. I'm changing the output to "success" and "fail" to see if I can get it working reliably.
Gabriel Echarte
·
·
I streamlined the prompt a little using our metaprompter. It's outputting Success or Fail now in a test duplicate column. Do you want to look through this and confirm it's what you wanted? Attaching explanation of issue below!
Gabriel Echarte
·
·
Why This Happened:
1. OpenAI Rate Limits ** Your API key has usage limits (requests per minute/day) ** Claygent makes multiple API calls per LinkedIn analysis ** When limits are hit, requests get queued or fail silently 2. Claygent Timeout Issues ** Each Claygent task has a timeout limit (usually 60-120 seconds) ** Profile scraping takes time, especially for complex profiles ** If analysis takes too long, Claygent stops and returns blank
How to Confirm It's Happening: Check OpenAI Usage: 1. Log into OpenAI dashboard 2. Go to Usage > Activity 3. Look for rate limit errors or high request volume 4. Check if you're hitting monthly/daily limits
How to Fix It:
Option 1: Reduce Batch Size ** Run 10-20 rows at a time instead of hundreds ** Wait 30-60 seconds between batches ** This prevents overwhelming OpenAI's rate limits Option 2: Upgrade OpenAI Tier ** Higher usage tiers get better rate limits ** Check your OpenAI tier status and consider upgrading Option 3: Simplify Prompt Further ** Our streamlined prompt still might be too complex
Hasnain R. (LeadsOps)
·
·
Hey Chris Balestras! It could be due to missing inputs a broken prompt formula issues or failed columns it depends on double-check those and re-run.