When
you and I have a conversation, the dialog is fluid. I might make a statement,
and then you might ask for clarification. When implementing a Virtual Assistant
(VA), the idea is to replicate (as much as possible) human interaction. The
implied objective is to pass the Turing Test, where you and I may not know the
difference between chatting with Artificial Intelligence (AI) vs. interacting
with a person. I agree. However,
we have to ask ourselves: What is the purpose
of the VA we are trying to create? If the goal is chatting to entertain, then
it should be architected to pass the Turing Test. But most of the time,
customers require a VA that must complete a task, or perform a specific
job such
as solving
a technical problem, or helping a user choose the best product based on a
specific need. In this case, efficiency may have to be prioritized over
human-like dialog.
This
is my view based on creating various types of VAs with very specific company
goals over the last 16 years. You can click here to contact
noHold to learn more -- or message me on Linkedin.
Cheers,
Diego
No comments:
Post a Comment