I've been experimenting with the Vambe Ads wizard, and for general tasks (pulling data, performing high-level analysis of creatives, ad sets, and campaigns), it's capable of generating good insights — but when it comes to discussing different ways to structure marketing campaigns, it falls very, very short; on several occasions, it has drawn false conclusions.
Example #1: While discussing an A/B test campaign I’m setting up, I asked it about the implications of data flowing from different Events Manager datasets. It told me that one of the datasets hadn’t had any data for days (which was false—I verified it directly in Events Manager).
Example #2: During that same session, I showed him a screenshot from EventsManager showing that data had been flowing recently—he warned me that the event type we’re generating wouldn’t work for my direct-to-WhatsApp campaign (LeadSubmitted), but that event is the one generated by BusinessMessaging when sending a Lead-type event, and it does work for campaigns sent directly to WhatsApp.
In my opinion, if they improve the model’s intelligence or fine-tune it for more nuanced scenarios, it could be useful in these cases
