I’ve just updated my claygent column settings to increase the max spend per run to run any of the cells that didn’t run initially, and it has rerun all the existing results and lost the results to ~15,000 rows. Is there any way to retrieve the results? https://app.clay.com/workspaces/27790/tables/t_q5YfsAutwoAE/views/gv_eX6c5ufk0COX
Hi Anjali, when I first added the column I would have asked it to run 10 rows and then all rows that haven’t run or that have errors. When I updated it to increase the max price, I asked it to run all rows that haven’t run or that have errors.
Got it, Joel S.. It's definitely strange that some of the Claygent responses are 'Yes' even when there is no site listed (in which case the only input to the AI would be the location). Can you try including the spa's name as an input along with the site. This might increase the accuracy more, and should you rerun, hopefully attain more valid results.Sorry for the inconvenience you faced, AI isn't always reliable and sometimes the language used with it is key. I hope adding a bit more specificity resolves some of your issues!
I have no issue with the results, the reason it could get a result is I was feeding it the Google Maps location. I was asking if it was possible to retrieve the results without rerunning ~16,000 rows, but I have now done so anyway. Can you please look into the logic that reruns an entire column so that something like increasing the max cost doesn’t cause the whole column to run again?
Hey Joel, sorry for the delay! I just read this, and unfortunately, it's a bit complicated to check for that. The thing is that if you change something in the prompt or settings, the API call to that provider would be considered a new request by their servers since it directly modifies the entire API structure with such a change.We do have some checks in place to ensure that if nothing changes or if the inputs remain the same/keep their exact structure, then nothing should re-run.
Added extra credits to your account for the situation though. :)
Thanks Arturo!