Hello all, I am CoFounder and CEO of XaiPient. We have built an app for B2B HubSpot users that leverages our innovations in Trustworthy AI models for event sequence data. We apply ideas from language modeling, such as Recurrent Neural Networks, self-attention and Transformers (the latter two underlie the famous GPT3, BERT and varients).
The app's learning module learns from all prospect histories in a HubSpot portal, and predicts each account's (relative) chance of becoming a customer (Closed Won) within some chosen window (default 90 days). It also shows attributions, i.e., how much each historical touchpoint contributed to today's score. These attributions are shown on specific account journeys, as well as in aggregate. The idea is to help marketing/sales teams better prioritize accounts for intensive sales accounts and allocate budget to the most effective channels and campaigns.
A recent article on our blog presents the context of the problem we solve: https://blog.xaipient.com
You can find our app by searching in the App Marketplace for XaiPient.