Linel, the AI Tutor
How I designed the personality, conversational behavior, and support boundaries of an AI tutor students actually wanted to talk to.
My Role
Conversation Design | Copywriting
Persona Development | Strategy
The Problem
We had an AI bot,
but nobody wanted to use it.
When GenAI hit, every education company scrambled to bolt an "AI tutor" onto their product. Varsity Tutors already had one almost nobody used it.
It could answer questions, but it couldn't respond to students. It had no read on urgency, frustration, or what a student actually needed in the moment. It had been treated as a feature, not a conversation.
The real opportunity was to design a tutor that could tell the difference between "I'm stuck" and "I'm struggling," and answer each one well.
My Part
I owned how Linel thinks and talks.
As part of a brand personality workshop, I ran an archetype study and worked on the scenario framework. I wrote the voice, tone, and messaging frameworks that define how Linel speaks. I also set the boundaries for how much emotional support Linel should provide.
We partnered with external agencies on visual design, motion, and prototyping. I directed the conversational layer across all of it.
We partnered with external agencies to explore every layer of the experience, from how Linel introduced itself to how it responded when a student got something wrong. The work covered its visual personality, conversational patterns, and the small interactions that make a product feel alive. The goal was a tutor that felt like a great teacher. Patient, curious, a little playful, never condescending.
The Approach
We weren't just deciding what the AI tutor could do. We were deciding what it would say to a stressed-out 13-year-old the night before a test. The work moved in five stages:
Define the conversation.
Then define the personality.
1. Personality workshops with stakeholders to define what the tutor should sound like and stand for.
2. An archetype study to find Linel's core character.
3. Scenario mapping to decide how Linel responds to the real moments students reach for help.
4. Voice, tone, and messaging frameworks to keep Linel consistent across every interaction.
5. Testing with middle and high school students to see whether the personality landed.
We partnered with external agencies to explore every layer of the experience, from how Linel introduced itself to how it responded when a student got something wrong. The work covered its visual personality, conversational patterns, and the small interactions that make a product feel alive. The goal was a tutor that felt like a great teacher. Patient, curious, a little playful, never condescending.
The Archetype
We weren't just deciding what the AI tutor could do. We were deciding what it would say to a stressed-out 13-year-old the night before a test. The work moved in five stages:
A little bit sage.
A little bit Caregiver
Look and Personality Explorations
The mock-ups showed how Linel would live across subjects, sessions, and moments of struggle. Each touchpoint was designed to feel less like a tool and more like a presence. Someone in your corner.
App Interaction Mock-ups
The Result
Validated by students.
Ready when Nerdy is.
The work earned full executive buy-in. Initial testing with middle and high school students was the most encouraging part: they responded to the personality, the clarity, and the fun. They wanted to keep talking to it.
The project was eventually shelved due to shifting priorities, but the messaging foundation, the voice, the personality, the framework, remains intact. It’s still ready to power Nerdy’s AI tutor when the moment is right.