Reading Rocket
AI-Powered Literacy Tool
Reading Rocket is an AI-power literacy tool that helps teachers assess and analyze K-2 student literacy needs.
Position
Product Designer
My Role
As the company’s sole product designer, I have a range of responsibilities including product research, product design, domain expertise, and partnership development.
Date
October 2023 - Present
Methods & Tools
User Interviews
Competitive Analysis
Site Analytics
Information Architecture
Prototype (Figma)
Usability Testing
Domain Expertise
Graphic Design
Table of Contents
Challenge & Solution | Context & Process | Review | User Insights
Ideation | Final Design | Additional Features | Usability Testing | Discovery
Challenge
When assessing the literacy needs of students, education professionals are faced with the following challenges:
Assessment accuracy
Time and energy to
assess students and analyze data
Personalized and engaging
literature
Solution
Reading Rocket aims to design an AI-powered literacy tool that helps teachers assess / analyze K-2 student literacy needs and generate personalized texts.
Context
Reading Rocket is an AI-powered literacy tool developed by an early-stage edtech startup. The company's technical founder, Ron Potok, launched Reading Rocket after becoming fascinated by watching his mother, a career educator, teach his niece to read. Ron holds a PhD in Physics from Harvard and has over 15 years of experience building high-performance data science products, including teaching computers how to read.
I joined the company as the sole product designer, building on an initial round of research and design. Collaborating closely with Ron and our team’s learning specialists, I designed over 7 new features that improved the tool's usability. To achieve this, I established partnerships with literacy organizations, conducted usability testing, analyzed research, and leveraged findings to design solutions. These design improvements made the tool more user-friendly and helped us better meet the needs of our end users—students and teachers.
Process
I grounded my approach in the ten steps of the product design process, while staying flexible and adjusting accordingly as we navigated the challenges and pivots involved in launching an MVP.
Review
I analyzed the research and designs provided by previous designers on the project and reviewed the requirements for the MVP. I then provided a first round of design critiques for two features: the read feature and the student dashboard.
Analyzing Research and Designs in Handoff
I reviewed the teams existing market research summary that establishes schools/districts as the main buyer, considers the needs of all those involved in teaching literacy, and addresses business goals and revenue streams.
Market Research
Our founder provided a competitive matrix that shows how Reading Rocket combines diagnostic monitoring with generative AI, and can adapt to any curriculum. In comparison, other companies either focus solely on AI diagnostics monitoring (ex. Ello, Edsoma and SoapBox Labs) or generative AI (ex. Amira, LitLab and Project Read).
Competitive Analysis
The conservative TAM was estimated at $1B in annual recurring revenue (ARR), with $450M ARR from schools and districts and $500M ARR from the consumer market for digital books.
Addressable Market
User Insights
I conducted interviews with over 30 education professionals, including school leaders, teachers, lawyers, publishers, designers, investors, and founders. During these interviews, I gained insights into the challenges education professionals face with literacy education, and also demoed our initial Reading Rocket designs to gather feedback.
I sourced interviewees from both my professional network and through cold outreach.
Interviews
RESEARCH FINDINGS
1
Demand for accurate child pronunciation recognition
2
Interest in personalized, engaging AI content
3
Safety and truthfulness in AI top of mind
4
Save teachers time and energy when it comes to assessing students and analyzing data
During interview research, I also launched a search for local literacy programs that might be interested in partnering with us or supporting usability testing. My search led to a collaboration with two literacy programs for usability testing and a co-authored grant, opening the door for a potential future partnership.
Ideation
To enhance Reading Rocket and prepare it for more in-depth usability testing with teachers, our team prioritized the following features based on our internal design critique and the insights gathered from interviews:
Efficient sign up / log in
High level preview of reading accuracy
User-friendly assessment creation
In order for teachers and students to access the Reading Rocket platform for usability testing, we needed to create an efficient sign up / log in process.
Key Finding: K-2 students and teachers require a time efficient and simple sign up / log in process
Solution: During our research we found that schools commonly use Clever as their go to single sign on method. My initial flows provided multiple sign up / log in options for students and teachers including Clever, Google as well as a Reading Rocket account. We also brainstormed ways to support teachers when it came to logging in students.
Design #1: Efficient sign up / log in
Gather Inspiration
I researched the sign-up and log-in flows of similar literacy tools and discovered different approaches to onboarding multiple user types. A key theme for student log-ins was allowing teachers to sign up first then provide an access code for students. This approach eliminates the need for young students to sign up themselves.
Mockup
I created mockups to explore how we could clearly communicate the phonics skill and concept selection process. The mockups visualized different approaches, including combining both into a single step using different designs or breaking them down into two distinct steps.
Final Design Decision
Ultimately, we chose to display the phonics concepts (e.g., CVC words) as buttons, making them visible and easy to find while fitting neatly in a few rows. Once a teacher selected a phonics concept, they could then refine their choice by selecting a phonics skill from a drop-down menu. We placed the skills in the drop-down menu because it was optional, and displaying too many at once could feel overwhelming and cluttered.
Information Architecture
I created an information architecture diagram to visualize the structure of two log in flows (teacher and student) and one sign up flow (teacher).
Mockups
I created mockups for the teacher sign-up/log-in process, including options to create an account with Reading Rocket or through Google/Clever.
Our design requirements included:
Two SSO options (Google and Clever) plus a Reading Rocket account option
User type: Teacher vs. Student
Name
Create / confirm password
Teacher display name (ex. Ms. K)
Email
School selection
Profile image
We needed to strategically plan when to request each piece of information to ensure user retention and minimize drop-off during the sign-up process.
I also created mockups for the teacher / student log in process which took into consideration ways to support student log in and how to provide two clear pathways for both teacher and student.
Key Finding: One key takeaway from our interviews was that K-2 teachers overwhelmed by the amount of data they had to analyze.
Solution: We designed a card that provides the high level data analysis of reading assessments at the top of the results page, allowing them to quickly and efficiently access the most important information at a glance. Teachers can click on the card to deep dive the data analysis.
Design #2: Preview of reading accuracy
Mockups
I created mockups for the data analysis cards, which featured key details like overall reading accuracy of the student assessment, the date, assessment type (phonics concept), and the book name. In design 1, I emphasized the most important information—the accuracy percentage—by making it large and easy to spot amidst other competing details.
Key Finding: Another key takeaway from our interviews was that K-2 teachers were unsure how to find an assessment that focused on the specific phonics concept or skill their students needed to practice. This issue stemmed from our initial book selection process, which was based on a numerical scale relative to grade level rather than specific skills. For example, teachers were prompted to select a book by entering a number: 1–10 for kindergarten (1= beginning of kindergarten, 10=end of kindergarten), 11–20 for 1st grade, and 21–30 for 2nd grade. Teachers were then required to search for relevant phonics skills amongst a selection of books. This lengthly process left teachers feeling confused about how to find the right assessment that best met their student's specific needs.
Solution: We designed an assessment creation process that had clear step-by-step instructions, the option to select a specific phonics concept and/or skill, followed by the ability to select a book associated with that phonics concept/skill.
Design #3: User-friendly assessment creation
Sorting phonics books by concept and skill
To improve the experience of selecting an appropriate reading assessment, I focused on sorting our books by phonics concepts (ex. CVC) and phonics skills (ex. short view u and short vowel a both fit under CVC). I sorted the books based on a scope and sequence I used during my time as a teacher in the classroom. I also collaborated with two education specialists on our team to determine what the most up to date scope and sequence is in relation to the science of reading.
Sketch
I sketched different ways for teachers to efficiently search for their desired skill including drop down search menus and buttons. We ultimately incorporated both designs as you’ll soon see.
Key Finding: When teachers launched the book, they weren’t sure if/when the assessment officially started recording.
Solution: I designed a cover page for the book and wrote a clear action statement on the start button to let teachers understand they will begin to read and record once they click the button. I also added a recording symbol to the top of the page to confirm recording is in action.
Design #4: Reading book record button
Final Design
The final design features a/an:
Efficient sign up / log in with two SSO options, including Google and Clever, and an option to create a Reading Rocket account
User-friendly assessment creation that sorts books by phonics concept and skill
High level preview of reading accuracy before deep diving the analysis
Additional Design Features
We created a student dashboard with two features:
Assessment: Students can independently access assessments shared by teachers under the “Show What You Know!” section.
Library: Students can independently access personalized books in their library.
Finding: Based on feedback from our advisors, we found that adding this feature to the MVP/MVT may not be advantageous at this stage. We decided to focus primarily on improving the teacher dashboard which allows teachers to create assessments, launch assessments (share the device with student so student can read), and analyze the assessment results.
Design #1: Student Dashboard
Usability Testing
We conducted usability testing with 20+ teachers and students. Our in house learning specialists used the tool with their students and teachers. To source additional participants, I reached out to a speech, language, and literacy organization in New York which resulted in four usability tests. Our team also shared the tool with friends and family.
I created step-by-step instructions to support our participants understanding of
our objectives and how to use Reading Rocket (see here).
Key Findings: Our key finding was the need for improved accuracy of our model. For example, teachers found that the analysis showed a mispronunciation when the pronunciation was correct.
Solution: We’re continuously improving the model by training it with new data collected through sharing the tool and conducting ongoing usability tests with students and teachers.
See my Kid Smart AI case study to learn how we continue to improve the pronunciation model.
It was nice to see the different ways he had errors in his reading, and we were able to discuss the times he mispronounced a word or where he corrected himself. It showed his weaknesses and the areas he did well, which was important when discussing the results with him…. Bennett came to my house yesterday and said, "I want to do another one of those tests!" He is so self-competitive and wants to get 100% on the tests. So, for us, this is going to be a great program for him.
-Nicole T., Director of Literacy Program
Discovery
Ongoing research revealed a demand for speech recognition services among education companies. We specifically targeted two distinct markets to assess which one was worth pursuing.
Replacement: Replacement speech recognition service for companies that currently use speech recognition
Add-on: Curriculum companies that do not currently offer speech recognition in their products
Check out the case study on our pivot to designing AI solutions for edtech companies with Kid Smart AI.