We've received several requests to improve our experimentation functionality:

* Support testing more than two variants - allow AI to create more variants

* Inform whether the test achieved statistical signifigance or not

* Test campaign on positions & triggers

* Test on more events, such as the copy coupon-code-event

* A UI more integrated with regular campaigns.

* Support converting an AB-test back to a campaign without having to publish the test and select a winner.

Do you have suggestions? Please comment below!