Edifius AI-Powered Software
Software company is in need of an onboarding flow that connects small business owners to their customized voice bot.
Scroll ↓
Challenge
Non-technical, small restaurant owners need an easy way to learn and train a voice bot taking phone orders.
Solution
Develop a purposeful onboarding flow with minimal friction for the busy pizzeria owner and their staff.
My Role and Contributions
Served as the project manager and ux researcher during this three-week design sprint
Conducted user interviews and usability tests at varying fidelity during design iterations
Fostered a working remote relationship with agency and stakeholders while guiding the team’s productivity through the agile development process
Research Methods
Stakeholder interviews, Competitive Testing, Heuristic Evaluation, Task Analysis, Journey Mapping, Usability Testing, NPS Surveys
Our Goals
Develop a quick, and purposeful onboarding flow for the non-techie pizza shop owner.
Educate our user on voice bot’s ability to engage their customers over the phone.
Draw in users by making the process as simple as possible.
Standing Up to the Competition: Direct Competitor Task Analysis
It was important to not only identify pain points in our current onboarding flow but also within our direct competitors to identify potential opportunities.
In order to avoid bias assumptions headed into user interviews, we first wanted to first conduct a task analysis of our direct competitors. Our limited access to users who fit our proto persona (the small business owner) was our first challenge, resulting in the use of interview proxies.
Competitors display shorter onboarding flows but do not offer access to the product.
Six participants were given the task of onboarding onto a competitor’s website in an attempt to identify:
Ease of onboarding
Speed of onboarding
Engagement with product
Awareness of product
Conducting User Interviews By Proxy
With the wealth of background information provided by stakeholders and the insights gained from our task analysis, we were ready to begin coordinating user interviews. Five participants were selected to serve as proxies, each specifically asked about customer phone interactions, phone call content and frequency, and the level of disruption to restaurant operations. In addition, participants were asked to share insights on how they believe the voice bot (Simba) would integrate into their previous workspace.
What users were saying…
“Can the voice bot do the job?”
Users Insights Revealed:
Curiosity around the voice bot’s ability to handle complex customer calls
Concern for the existing customer experience
A high expectation around ease of learning for all employees
A need for simple integration with existing restaurant software (POS systems, food delivery apps, etc)
The desire to learn more about voice bot’s ability to serve multilingual customers
Meet Our User Persona Sean
Insights gained from user interviews allowed us to develop a user personal with more focused goals, frustrations, and expectations.
User persona Sean.
Identifying the Core Problem and Potential Solution for Persona Sean
I created this journey map for our team based on user insights and team assumptions to gain a better understanding of Sean’s emotional journey onboarding Simba.
Now that we have a better understanding of user needs, how do we address concerns while highlighting Simba’s key features and functionality?
Building Trust in Simba Through a Frictionless Onboarding Flow
Objective: To rank design features for Simba’s onboarding flow. These features are an accumulated list of design patterns or tools uncovered through our previous task analysis.
Potential features ranked under the MoSCoW method.
All potential features to include in the grayscale prototype were ranked under the MoSCoW method. For those unfamiliar, MoSCoW is a method of classification that allows us to view these features with the perspective of what is critical to design versus what is not. Each ranking resulted in features that enhance the onboarding flow as compared to competitors.
Grayscale Usability Testing Revealed: Users Needed Specifics
We wanted to build the first prototype at a very low fidelity to explore different paths for the user while breaking down the existing long-form onboarding into smaller, more palatable steps.
Users expressed confused about where to go once they arrived at the three module page.
Users sought clarity around language being used throughout the prototype. “I’m finding it difficult distinguishing the difference between Simba, virtual assistants, and virtual employees”
Lo-fi Usability Testing Insights:
Users needed more specifics regarding the voice bot’s core features. “I need to get to the core value of what Simba can do for my business sooner.”
In addition, users thought the Configure Me module referred to voice bot setup rather than voice bot registration.
Users wanted clarity about why they had to assign human features, such as gender and a profile photo, to something not human. “Why does Simba need a profile picture if I’m talking to it on the phone?”
Users complained about having to click too much in the learn about me path
Design Opportunities
Making copy more conversational with clear instructions to guide the user.
Remove bot profile and incorporate required questions throughout flow.
Iterating From Lo-Fi to Mid-Fi: Pivoting The Focus to Payment Plans and 14-Day Trial
In this next design iteration, we shifted the focus to the 14-day trial and various payment plans.
Imagine sitting through this screen for five minutes as a busy restaurant owner.
Usability testing of the mid-fi prototype revealed:
Further confusion around payment plans and a need for clarity around conditions of the free trial
Difficulty understanding the available options and if the key features presented were for one specific plan or all plans
Users didn’t understand what Simba was
Users felt that five minutes was too long to wait for a phone number to generate.
Users also communicated too many CTA’s and not enough information on the home screen.
3 out of 4 users had questions around the transfer phone number and message.
Users still appeared apprehensive about selecting a payment plan so early into the process, but responded positively to the quick staging number set-up.
Design Opportunities
Moving payment plan selection toward the end of the flow within the dashboard
Make CTAs on home page more focused along with streamlining Simba’s features above the fold
Eliminating the mention of the transfer number within the onboarding process
Give users the opportunity to move into the dashboard as Simba generates
Iterating From Mid-Fi to Hi-Fi: Piecing the Flow Together
By testing the prototype, we look to generate qualitative and quantitative key performance indicators (KPIs) that allow us to measure the effectiveness of the new onboarding flow. A net promoter score (NPS) survey is implemented at the end of the user test to establish a metric of success as we iterate throughout the next stages.
High fidelity usability testing in progress.
Four key insights from our proxies in the interviews:
Users wanted direct and relatable examples of the key features Simba could provide them, as compared to its technical capabilities
For users to take the leap of faith and purchase a plan, we need to increase their exposure, familiarity and reliance on Simba
Users wanted more clarity on the payment plan, specifically why they would be billed per call, and the ambiguities that surround this
Users wanted a greater range of accessibility options, from a range of improving the UI legibility to providing a full-audio on-boarding experience
Next Steps and Recommendations
Explore the range of accessibility tools that may be required by a potential portion of our users
Consult branding and identity at Edifius HQ to bridge the educational marketing content on the onboarding with the existing visual identity
Explore additional (financial) incentives in the form of free calls to strengthen user’s initial onboarding momentum
Add more interactions and feedback for the user via mascot
Add a ‘menu builder’ step to onboarding flow
Build out flows for help/guide documentation paths
Consider gamifying the onboarding flow