Lossless Voice
Simplifying AI Assistant Management Experience for Restaurant Owners
Project
Overview
Lossless Voice is an AI voice assistant that handles calls and drive-thru orders for restaurants. The dashboard enables business owners to view and analyze their business insights and manage and improve their AI Voice Assistant.
Problem to solve
Other AI assistant products require a long calls with the sales team to get an account set up. Our goal was to design a platform that simplifies this process and simultaneously presents business analytics collected by the AI assistant.
Goal
Design a platform that empowers restaurant owners to set up and manage their AI voice assistant online easily.
Challenge
As the only designer on the team, I was responsible for product management. I made user-centric design decisions with consideration of engineering and development.
To Kickoff the Project
As a brand new startup we had very tight deadlines. I struggled to find time for in-depth user research, especially as I was the only person working on design.
How I collected insights under time constraint?
Conducting Competitive Analysis
I compared various AI assistant registration processes and identified key elements that enhanced and hindered the usability.
Frequent Usability Testing
I conducted frequent usability testing internally and externally throughout the design and development processes to ensure the design was aligned with users needs.
Constant Communication with Team
I communicated daily with my team to ensure my design was aligned with the development, and that all elements could realistically be built on our tight schedule.
Planing the Product
How might we make it possible for business to easily sign up and deploy an AI assistant by themselves.
Product Roadmap to Design Plan
As the only employee at the time of project kickoff, I studied the full product roadmap by discussing it with the CEO. Based on these discussions, I created user flows and drafted the initial design plan. This helped me to prioritize my tasks.
Teamwork
As the team expanded I introduced the user flows. We revised the priority list for the MVP and decided on the features that would be included.
Some features were easy to prototype but heavy work for the development team so we push these back to the alpha and beta stages.
Daily Design Reviews
I created wireframes based on the task flows. We reviewed whether my sketches aligned with the vision of the product.
Feature Highlights
A collection of unique features, many of which were conceived from user insights.
Easy setup process
To reduce the hurtle for restaurant owners, we’ve created a intuitive setup process allowing them to subscribe, try, setup and manage AI Assistant all through online.
Shadowing Service
With the poor reputation of voice assistants in the current market, I designed a mandatory shadowing service that analyzes businesses’ phone calls before the AI is deployed to earn business owners’ trust.
Smart Ads
I designed a unique way to easily create and deploy AI generated advertisements within phone calls: Smart Ads, which creates ads for the user. It requires minimal input from the user, the AI generates ads, runs ads on target users, and provides metrics of the ad campaign.
LOSSLINK
To improve restaurant customers’ experience with the AI Voice System, LOSSLINK integrates with small businesses’ online systems, automating order confirmations and providing live updates.
Introducing “Foodinality”
We’ve designed “Foodinality” to offer restaurant customers food personality through playful ingredient analysis on the LOSSLINK confirmation page, sparking positivity and a memorable connection between restautants and their customers.
Testing and Iterations
Since we were not able to collect many research insights at the begiining of the product design stage, testing and iteration loops are critical for us to ensure the design is align with our users needs.
Usability Test Guidline
To collect more data points, I asked my team to conduct user testing. I created detailed guidelines on how to conduct usability tests to ensure we would collect high-quality data.
Test with Users
I ran usability tests with users who are not restaurant owners and received useful insights to integrate into the MVP version.
Testing in the field
To determine how real users felt about LOSSLINK and Foodinality, we went into the field to test these features with 20 users.
Due to the confidentiality, I will not be able to share any media about Usability Test.
One more thing...
Because I didn’t have experience building an entire product and there was no senior professional on the team, I was unsure if the steps I was taking were correct, I was lacking confidence. After talking through my doubts with my mentor I learned that no design process in real life is every perfect.
What did I learn?
Stay Hungry, Stay Foolish. I keep this in mind throughout all my designs and focus on learning from the talented people around me.
A Fluent Usability Test
A usability test on wireframes and prototypes that contain more detail and correct copy provided me with more valuable insight as the test flow was less interrupted. When the prototype’s copy is not accurate, users are confused and need explanation and guidance which could skew the test results.
Advocating
As the only product designer on the team, not only did I need to stand for my design decisions, but also introduce my team to the importance and value of human-centered design.