We want to put our experience of AI usage into our products.
The 1st implementation was predicting identical leads where you define the segment depends on your conversion stage or CRM push.
We are currently working on Supervised training on Aritic-AI and will tell when we can start working on unsupervised learning to define the campaigns flow for you.
Any of the activity done on the leads or contacts sent to ML engine, where ML engine, keeps track of history required for decision, it just matches the behavior and provides the output leads category on the learned segment.
It was the initial starting of our AI-ML model, which works very well in the range of 67% to 92% accuracy. Since it was the 1st steps into ML of Aritic-AI, so we were on the right track.
Then suddenly one of our customers told us – “I have enabled a segment – named (Not a hot prospect” and “Prospect for future”). Even after two months and 1650+ contacts/lead, your AI is not giving any similar lead’s notification, so that our Marketing and Sales team will avoid those Contacts/Lead/Prospect.
Now, this is the 1st real problem posted to Aritic-AI and was trained only to act on the behavior or action taken by contacts/leads/prospects, in these cases prospects were not doing any activity. It has become very hard for Aritic-AI to give similar leads.
We are started working on increasing the decision forest and adding more features sets with the capability of “Label” features to contacts, segment, and campaigns.
I will keep posting on this page on Aritic-AI improvement and challenges.