AI Reader in Smartbook
For students who have limited at-home access to paid tutoring, they're often left to learn tricky concepts alone. At the start of 2026, I designed AI Reader in Smartbook, which gives underserved students personalized learning support directly within their homework. Today, it accounts for 91% of AI Reader's usage in McGraw Hill products.
Louis Nejame – AI Product Owner
Ethan has been the secret weapon for me to getting projects off the ground/unstuck etc. He's on our Design team and brings the forward thinking design we desperately need, and has skipped a lot of process for me to get us great mockups.
In addition, Ethan just rocks so he's a good guy to know
Highlights
Feature adoption
52M interactions from 2.3M+ students since launch
Students using AI Reader
Spend 2.12x more time on their assignments
Role
Product Designer, Vision
Team
AI Product Owner, Project Manager, 2 Developers
Timeline
2 months
Context
AI Reader is a gen-AI tool with simplified explanations for any selected text inside eBook. Smartbook is an adaptive assignment that adds concept-based learning onto eBooks by pointing students towards reading for what they don't yet know.
Today's reality: ChatGPT doesn't cover learning science
Over 60% of students use AI to look up the answer. By shortcutting learning this way, they might achieve basic recall for what they learned, but these same students show less ability at understanding, analyzing or applying those concepts.
McGraw Hill's answer: AI Reader launched in eBook
AI Reader in eBook launched in 2025, and within 3 months, 19,980 students were actively using AI Reader during their assigned eBook readings.
Smartbook has 3x as much usage, so the AI product owner and I started to build a case for AI Reader in Smartbook, potentially reaching an additional 3.3M+ more students.
How might we put a successful feature like AI Reader into a feature-robust assignment like Smartbook?
I wanted to send students straight into deeply learning concepts
Currently, students will read a Smartbook chapter, answer a few questions to see what they learned, and then reread about concepts they got wrong. AI Reader naturally found a home to support to students once they got a question wrong.
I tried placing AI Reader directly within a question, side-by-side to a question, and even turning AI Reader into an AI prompt after a wrongly answered question.
Should AI Reader be first or second next to authored content?
Embedded directly in questions
Students almost always entirely skipped authored content. This was really concerning when a student took a hallucination as the truth.
Secondary to eBook content
I interviewed two authors, too, and they felt completely replaced. Trust would be broken if we went showed AI content first.
Since Smartbook was much more complex, should the form of AI Reader be more digestible?
I tested AI Reader as a popup for each concept with students.
But students quickly pointed out the loss of persistent history, and we knew this would be more troublesome for our 74% of students using Smartbook with their smart phones.
Two additional, hidden gems appeared through our constant feedback loop
Students gave overwhelmingly positive responses when I tried out AI Reader being openable directly beside every concept. It became our main selling point.
During interviews, I had noticed students switching between the reading and questions tabs constantly when they forgot what they were doing. Having the question stem show was a top-loved idea that made it's way into the product.
Post launch, AI Reader in Smartbook brings deeper learning to the students looking for it
52M interactions from 2.3M+ unique users, 91% using Smartbook
Students with usage spend 112% more time on the platform than non-AI users
No signs of systematic cheating. (-5.1pp first-tries correct)
This project's process was constantly changing, and it produced fast, chaotic ideas
On paper, we did two testing rounds; n=6 and n=7 with students. In reality though, we changed designs, tried new directions, and shared different ideas after almost every call we had, and it was very much the three of us trying a million things out.
Reflection
Working in such a minimal, blue-sky space showed me the value in tight, consistent feedback loops. I was in calls with students and teachers every day, showing design ideas, and furiously trying new ideas based on our conversations.
This was my 10% project at the time, something I took on outside of work hours. The things I learned and the connections I made are some of my most valuable ones yet.