Overview
Using human-system relational modeling, virtual assistant best practices and business requirements I lead the team in planning our next year priorities based not just on business requirements but in prioritization to build trust and value to the end user.
Opportunity
The pipeline for requirements was business driven, fighting for differentiation in the marketplace without starting with a baseline table stakes tool. Our team was focused on fixing common problems with getting stuck in a poorly authored virtual assistant, but in a vacuum.
Business goals
Increase user adoption and conversion rates
Design Strategy
Identify the cross-section of business identified features and human values to develop a system-human relationship and increase the value of web chat in the eyes of the end user.
Relational development practices
Back in the days of the physical office…
Knapp Relational Model
Initiating/Purpose
Identify the assistant as a non-human entity
Demonstrate value immediately
Provide onboarding
Experimenting
Keep conversation logic loose
Automatically turn unknowns to knowns (disambiguation, more options)
Intensifying
Redirect, refocus, reprompt - in the event the user diverts from the main subject
Sub actions (multi-turn)- a standard set of steps that live within the primary action, grouped to be distinct from the main conversation
Jumping actions- redirection of task based on end user changing focus, due to user or error of bot.
Integrating
The assistant should guide the conversation
Provide creative fallback messages - when possible assistants can support assistants
Allow the assistant to bring in information from the website, news sources or other data bases
Bonding
Allow for personalization
Don't show the user information they don't understand
Error messages must take responsibility and clarify the issues
Let the user report unique bugs in the system
Inform the user when the system is down
Allow the user to reach a live, human agent
Matching best practices with the user journey
Next step was to look at a session life cycle of the user when interacting with the assistant. Questions I had were:
At what points might the user hit a situation where these best practices come into action?
What best practices could be smoothed over with components and technology vs the actual content developed by our clients?
What would a roadmap look like for supporting these best practices from a web chat point of view?
User - assistant interaction flow
Alignment through visual collaboration
After the initial proposal of this best practice approach and highlighting areas of opportunity through the user flow shown above, I ran a working session with my lead offering manager and front end developer where we looked at what was in progress and prioritization of what was to come next. This working session influenced our 2021 strategy planning.
