
Member-only story
H2H vs. M2H Conversational Modeling: Part 2 — An example
This post is Part 2 of a multi-part series of posts on the topic: When designing a voice first conversational experience, should the designer strive to emulate the human-to-human (H2H) model or should they adopt a different model, a machine to human one (M2H), in their quest to deliver a highly usable user interface.
Part 1 of the series attempts to identify and articulate some rules and guidelines for designing a highly usable voice user experience based on the M2H design philosophy.
Part 3 provides a running dialog from the many contributors who have pitched in with their thoughts, suggestions, criticism, and observations.
All of the above articles mentioned, and this article here, are a work in progress. So feel free to provide your feedback either directly in this document or by emailing me at ahmed.bouzid@gmail.com.
Dialog Context/Assumptions
- Company Widget Factory is an ecommerce service/website.
- User is a customer of Widget Factory.
- User has their account linked with the assistant. So the assistant knows who they are and what they have purchased, etc.
- User has interacted with the assistant before.
- From the data we have gleaned, we know that in general, 80% of order status requests are about the latest order placed,
- For this particular user, it’s more like 95%
- For this particular user, when they reorder, it’s 70% of the time an item other than the most recently ordered.
H2H Model
_________________________________________________________________
(MA-1) ASSISTANT: Welcome to the Widget Factory. Great to have you back, Ahmed. I’m here to help you look up the status of an order, re-order an item that you recently ordered, or speak with one of our Widget specialists. So, what would you like to do?
(MA-2) HUMAN: Give me the status of an order.
(MA-3) ASSISTANT: Status of an order. Do you want the status of the latest order or a specific order?
(MA-4) HUMAN: The latest order.