Using GPT for Column-Wise Analysis and Segmentation in Data

Is there a way to have GPT or another LLM look at a whole column (vs a cell) and run analysis on the column? Usecase - I want to have GPT look at my output for a ‘company analysis’ claygent run and then take all of the rows and define 5-10 standardized Industry categories and Service Categories based on the analysis. Then on a per row basis, set the Industry and Service category TLDR - GPT for fine grained segmentation Is this possible?

5 comments

Bruno R.
·
·
Hi Sean, thanks for reaching out. Taking a look at this now!
Sean W.
·
·
Btw - did this successfully by exporting --> running this through chat --> pasting in the result, but it was messy (especially the pasting part…) Tried to do this again and I’m running into errors uploading to ChatGPT - maybe an issue with characters (unicode?) in the big sheet I was trying to upload If I could just do this in Clay that would 🧑‍🍳 🤌
Bruno R.
·
·
Unfortunately, the token reference logic operates at the row level, meaning you cannot reference cells across rows. To perform the analysis you suggested, you would need to download your table as a CSV file and run your categorization analysis using an external tool. This row-level reference limitation restricts your ability to access or analyze values across an entire column.
1
Bruno R.
·
·
Yes this would be the workaround that I'd suggest, I will note your feedback and relay it to our product team, I think it makes total sense to have a chance to reference cells across rows!
Channeled
APP
·
·
This thread was picked up by our in-app web widget and will no longer sync to Slack. If you are the original poster, you can continue this conversation by logging into https://app.clay.com and clicking "Support" in the sidebar. If you're not the original poster and require help from support, please post in 02 Support.