Thanks for reaching out @asivakumar That’s an interesting use case. Were you able to make progress? With Amplitude you can group users with a certain property value that was recently introduced. Here’s more on behavioral cohorts. And here’s a peer thread that might help! 

Hi, Thanks for getting back to me. The only approach I could see is to set the cohort to sync with User Profile API and then have my BE call the amplitude APIs the set / unset the tag. Was hoping to avoid the BE step. 

Would you share that URL on the feature that was recently introduced once again ? It’s looping back to this conversation. Also, how can I create a peer thread ?

Thanks again. 

Thank you @asivakumar. Please keep us posted as you make progress. You can message other Amplitude Community members by visiting their profile and post questions to our forums anytime! 

Thank you @Amplitude Admin . Do you think the Govern Add-on can be used for the use-case above ? 

Hi @asivakumar 

Can you elaborate on what you’d possibly model the fraudulent behavior on?

If it’s a combination of some event count only, you could try using Computed Property.

If it’s based on a value combination of user and event property, then you try deriving this tag using the Derived Properties feature.

If neither of the above options are suitable, then your cohort approach+user profile API call approach seems the best one for now.

Also, if you are sending data to Amplitude via a data warehouse, you can try simulating the fraud behavior in SQL and then set the is_fraud user property on the user before sending data to Amplitude.

HI @Saish Redkar Yes, Computed Property definitely helps. I’m having some trouble saving the computed result as a user property, but I guess it just takes a while to sync. 

Derived Property will also be very helpful - unfortunately, I don’t have the govern add on yet