Hey Chris,

Thanks for the submission. There is a lot of detail on what we do here, but the community is probably not a good place to go through laborious detail on this topic, but I'll address broad strokes below and give you some pointers to other info (see the youtube link at the bottom and also the video snippets on our product pages). If you have an account rep, arrange a discussion on the topic. If you don't, let us know and we'll do some introductions!

Broad strokes (including short demo videos): https://www1.pega.com/products/pega-platform/intelligent-virtual-assistant 

Our Intelligent Virtual Assistant (IVA) platform allows for work to be specialized for different user communities including end-customers, including quite complex work that certainly the vast majority of chatbots don't do today. So, that complex service case that you might have built for internal use - you can streamline, add question and answer prompts to, and now deliver on a range of different channels, including messenger, web chatbots or even (through Alexa etc) voice. It comes with its own NLP capabilities for disambiguation, or you can use the technology that comes with the platform of delivery (in the case of Alexa etc). It is able to react and respond to interrupt questions, and able to identify when work can be done anonymously, or when a customer needs to confirm their identity before progressing forward. 

Customer Service adds the ability to access knowledge when working through a workflow, as well as supporting escalation into the contact center, when a particularly difficult question requires it, or if the end-user asks for it. When an escalation takes place, we transfer the text of the interaction to date and also any unfinished work, which we assign to the Customer Service Rep (CSR). CSRs can also see work completed and have a view of previous chatbot sessions. Customers can continue working with the chatbot after completing a chat escalation discussion. 

Hope that helps!

https://www.youtube.com/watch?v=2YJI9jdPZOE