Elizabeth Rennie
AI Instructor Assistant
User Experience, Business Strategy, Generative AI
Overview
The AI Instructor Assistant was designed to help instructors quickly evaluate whether a textbook aligns with their teaching goals, while introducing generative AI in a low-stakes, exploratory context. Positioned as a knowledgeable, AI-powered guide, the tool aimed to reduce adoption time and support decision-making without requiring direct support contact.
As the lead UX Designer, I partnered with researchers, PMs, and developers to define the assistant’s purpose, design high-fidelity prototypes, and lead usability testing. We identified key risks related to discoverability and user expectations—many users mistook the AI for a basic chatbot. By iterating on interaction patterns and copy, I helped reposition the assistant as a clear, valuable tool that better aligned with instructors’ needs.
🎯 Tasks & Goals
- Enable instructors to experiment with leveraging AI to help with tasks like syllabus creation, summarizing content, curriculum mapping, and evaluating narrative content.
- Introduce our novice GenAI Instructors to the potential functionality and benefits in a low-stakes decision moment.
🎨 Design Deliverables
- High-fidelity mockups and prototypes
- Usability tests and findings
🤝 Stakeholders & Collaboration
- Product Leadership
- Product Managers
- Technical Product Managers
- Development Teams
- User Experience Researchers
📊 Key Takeaways & Results
- Title alignment was top of mind
100% of users asked how the title aligned with their learning objectives. - AI interest is growing
Most instructors agreed generative AI will play an increasingly important role in higher education. - Curiosity, not clarity
Early usability testing showed instructors were interested in trying the assistant, but many misunderstood its purpose. - Discoverability issues were significant
80% of testers didn’t notice the launcher until prompted. - AI was mistaken for support chat
Users assumed it was a standard help bot, not a generative tool.
💼 Business Need & Context
As generative AI tools rapidly entered the mainstream, leadership saw an opportunity to explore how AI could streamline the course planning process for higher education instructors. One of the most time-consuming tasks for instructors is evaluating whether a textbook aligns with their course goals, often requiring outreach to support teams, in-depth content reviews, or trial-and-error.
Business Goals:
-
Reduce friction in the textbook evaluation process
-
Position Cengage as an early leader in AI-powered educational tools
-
Support sales enablement by reducing dependency on human representatives during low-stakes exploration
-
Increase adoption by offering immediate, intelligent guidance at the point of interest
To meet these goals, we needed to design an assistant that would be intuitive enough for first-time users, aligned with instructors’ expectations, and able to communicate its value clearly without being mistaken for a support bot. This required thoughtful UX framing, clear copy, and early usability testing to ensure instructors understood both the tool’s purpose and how to interact with it.
🖼️ High Fidelity
Launch Call-out
The product details page provides users currently browsing titles or adopting a new title a chance to evaluate with the AI assistant.
iFrame AI Assistant Chat
AI-powered assistant embedded in the Cengage Instructor Center, designed to help instructors evaluate if a textbook is a good fit for their course.
📐 Research & Testing
The UX Researcher and I shared concerns about the assistant’s discoverability and users’ potential misinterpretation of its purpose, specifically, that it might be mistaken for a generic support chatbot rather than an AI-driven tool. By raising these flags early, we were able to advocate for and conduct a small-sample usability test to validate our assumptions and uncover opportunities for clarification.
Primary Goal
Identify any major usability issues to address pre-launch. Major usability issues can be defined as anything that could impact the user’s ability to engage with the AI bot, like issues with discoverability or accessibility.
Secondary Goal
Gather some attitudinal and directional data on how instructors feel about our use case for AI as an assistant when evaluating a specific title.
Methodology
30 minutes
Moderated Usability Sessions
5 Higher Education Instructors
Figma Prototype
📌 Key Findings
Users did not notice the Launcher
80% of testers did not see the AI launcher and needed to be prompted to continue reviewing the page to spot it.
Users don’t think the assistant is AI
They think it’s a regular chatbot which they don’t value. Phrases in the copy create a mismatch of expectations.
“Real-time support” gets misinterpreted as human support.
The learning curve of AI intimidates instructors.
They want to interact with the Chatbot but worry, they won’t know how to engage it to yield useful responses.
🗣️ User Quotes
“It’s going to give me things that aren’t helpful.”
“Waste of time”
“I don’t want to have to like teach it what I’m looking for.”
🗂️ Summary Overview
The AI Instructor Assistant was designed to streamline course planning by helping instructors quickly evaluate whether a textbook aligns with their teaching goals. Positioned as a knowledgeable, AI-powered guide, the tool aimed to reduce adoption time while also introducing instructors to generative AI in a low-stakes, exploratory context.
As the UX Designer on this project, I collaborated closely with researchers, PMs, and developers to bring this new experience to life. Early in the process, we identified critical risks around discoverability and user expectations—many instructors either didn’t notice the tool or mistook it for a basic support chatbot. By advocating for and conducting usability testing, we validated these concerns and used the findings to shape clearer UI patterns and copy direction. The result was an interface designed to educate, guide, and empower instructors while demystifying AI’s role in their workflow.