Seeking Success Stories with Claygent for Improved Prompt Results

Has anyone had really good success with Claygent that would want to maybe chat with me? I ran two different prompts to find the URL of a company based on a few (minimal) inputs. My second prompt used a better, cleaner AI model that returned a more hygienic (data) response, but the first prompt/model found more actual results. I need to find a way to get the best of both worlds and I've used a lot of credits in trial and error here on the prompts.

17 comments

Muhammad S.
·
·
I would use serper.dev to perform a search query on google with the inputs I have and then pass all the 10 search results of google to Chatgpt along with the context about the company I already have and have it pick the best domain from the google searches. This method is very accurate.
Kimberley C.
·
·
Is that one of the enrichments I can use inside Clay already?
Muhammad S.
·
·
Nope, but very easy to integrate within clay + very cheap. You can also do a google search within Clay by using built in google search function but it charges 1 credit per row.
Kimberley C.
·
·
Do you work there
Muhammad S.
·
·
work where?
Kimberley C.
·
·
serper
Muhammad S.
·
·
not yet,
Kimberley C.
·
·
Ok thx for the quick reply + tips, will take a look
Ariful A.
·
·
Muhammad S. what are you using for this to automate? make.com or something else?
Muhammad S.
·
·
clay is fine - you can integrate serper.dev in clay to do that
Ariful A.
·
·
What am I doing wrong with the body? It shows "Failed to parse body input"
Ariful A.
·
·
Muhammad S. any thoughts?
Muhammad S.
·
·
you can not add a text + a dynamic column within the body directly
Muhammad S.
·
·
create a formula column and add the text + dynamic column there
Muhammad S.
·
·
then pass that formula column to this body
Ariful A.
·
·
Sounds good. Thanks brother 🙏
Dan K.
·
·
I tested this on Muhammad S. advice a couple of weeks back and it upped the accuracy for UK oriented corp sites so thumbs up from me. 👍
👍1