Hey, I am doing lead scoring based on account research. Sometimes it researches 2 prospects that work at the same account, but I get 2 different scores. Did someone else experience this before?
Hey David, bunch of different ways you can implement scoring. Which method did you use? I’m building scoring into may of the systems I build and don’t have the issue. Happy to help.
I used Claygent. Basically I am doing this for a client and Claygent should research if the Account hosts any events at their location. This is one example:
I don't know why the output and confidence is different for the same account. Sometimes, if I run them again, I get different outputs. Could it maybe be that the prompt is too long and Claygent gets confused?
got it. I recommend another mechanism, That’s asking AI to make a lot of individual ’judgement. It’s not how you’d want to implement lead scoring at any scale. A way you can use for scoring is to build individual Claygents answering one specific question and add the results of that into a formula. You won’t get 100% perfect results (impossible) since the claygents can get things wrong depending on the data and how they are built, but a much more valid score in the end. Is this helpful for you?