Early Insights on Conversation Starters at Nectir

‍This blog post is a guest feature by Nectir’s Data Science Intern, Jonathan Trinh. Jonathan is currently pursuing an M.S. in Education Data Science at Stanford University. Before joining Nectir, he was an Economic Consultant, a Marine Corps Infantry Officer, and a three-year AP Computer Science and Mathematics educator with Teach For America in San José, California. 

For many students, facing an AI tool with a blank input box creates an unnecessary obstacle: Students aren’t sure where to begin, what kinds of questions to ask to support their learning, or what the AI Assistant can actually help them with.

At Nectir, we’ve been exploring a simple idea: what happens if students are given examples of productive ways to interact with AI before they even begin typing?

That question led to the development of Conversation Starters, instructor-curated prompts embedded directly into Nectir AI Course Assistants.

Over nearly two years, Conversation Starters have been rolled out across colleges, courses, and AI Assistants throughout the Nectir AI platform. Along the way, we’ve begun studying how students and instructors actually use them and what kinds of engagement patterns emerge.

While our ongoing causal research is still in progress, early descriptive patterns suggest that Conversation Starters may meaningfully shape how students engage with AI learning tools.

What Happens When Students Aren’t Faced With a Blank Prompt?

Conversation Starters are pre-written prompts that appear beneath the input box for an AI Course Assistant.

Instead of asking students to start from scratch, teachers and instructors can provide examples of productive ways to engage with the Assistant, such as:

• “Help me study for the midterm.”
• “Quiz me on this chapter.”
• “Walk me through this concept step-by-step.”
• “Create practice questions.”
• “Explain this topic in simpler terms.”

Before Conversation Starters, students were presented with a blank chat interface and had to independently decide how to begin interacting with the Assistant.

After the feature was introduced, students could immediately click into guided prompts or use them as inspiration for their own questions.

This may seem like a small interface change, but it reflects an important design question in education: How much does the interface itself shape the way students learn with AI?

‍

The Theory Behind Conversation Starters

Conversation Starters do not change the underlying AI model. Instead, they change the way students are introduced to the tool and the kinds of interactions the system implicitly encourages.

We designed Conversation Starters around three simple ideas:

• Reduce any obstacles to getting started.
• Model productive ways to use AI for learning.
• Give teachers and instructors more influence over how students engage with AI.

Conversation Starters do more than help students begin. They also shape the first impression students receive about what the AI Assistant is for. A prompt like "Quiz me on this chapter" encourages a different kind of interaction than "Give me the answer." By surfacing examples centered on productive learning behaviors, Conversation Starters are designed to gently nudge students toward more active forms of learning.

By providing examples of curated interactions, the feature enables instructors to precisely tailor and support the student experience.

‍

Conversation Starter Adoption

Conversation Starters were gradually adopted across institutions, courses, and Nectir AI Assistants throughout the Nectir platform. Since the feature's release in mid-2024, the number of Assistants has grown to more than 3,400, and nearly half (1,639) now include instructor-created Conversation Starters.

As availability expanded, usage grew alongside it. Students have now initiated over 20,000 conversations by clicking a Conversation Starter, and in a typical month, starters account for roughly 5–13% of all conversations on the platform.

‍

Students and Instructors Gravitate Toward Study-Oriented Prompts

One of the clearest patterns in the data is that both instructors and students gravitate toward prompts centered on studying, practice, and review.

The most-used Conversation Starter across the platform is simply “Help me study”, with over 4,000 student uses, more than triple the next most popular prompt. The rest of the top five tell the same story: "Walk me through the material," "Help me review for my midterm," "Quiz me on this course," and "Give me practice questions." Looking across all starter-initiated conversations by keyword, study-related themes dominate.

This alignment between instructor-created prompts and student usage is the most important pattern here. Teachers and instructors are encoding their educational priorities of studying, practice, and review directly into the interface, and students are taking them up by the thousands.

Conversation Starters appear to serve as a pedagogical channel, and they raise an intriguing possibility: that consistently framing students' AI interactions around practice and review may nudge them toward AI-supported self-study and influence how they choose to engage with these tools. Testing that causal link is a central aim of our ongoing research.

‍

What We’re Investigating Next

The patterns presented here are descriptive, but they raise several important questions that we are actively exploring.

First, what is the causal impact of Conversation Starters on student outcomes?

• Do students learn more when guided prompts are available? 
• Do Conversation Starters influence measures such as engagement, course performance, or long-term academic achievement? 
• And are some Conversation Starters more effective than others at improving student outcomes?

Second, how do Conversation Starters shape students' conversations with AI?

• Can we use natural language processing and unsupervised machine learning techniques to identify common patterns of student-AI use?
• How do different prompts influence the types of conversations that emerge?

Together, these lines of research aim to deepen understanding of how AI can support learning at scale.

‍

An Intentional Start for AI-Supported Learning

Since their introduction, Conversation Starters have been used to initiate more than 20,000 student–AI conversations across the Nectir platform.

What makes this particularly interesting is that Conversation Starters are more than just a convenience feature. They represent a new way for teachers and instructors to influence how students engage with AI. Rather than leaving every interaction to begin from a blank prompt box, educators can provide intentional starting points that reflect their course goals, teaching philosophy, and expectations for learning.

The early patterns we observe suggest that both instructors and students are embracing this opportunity. Study-oriented prompts dominate both creation and usage, pointing to a shared interest in using AI as a tool for learning, review, and academic support.

As AI becomes increasingly embedded in education, the question is no longer whether students will use these tools, but how they will use them. Conversation Starters offer one promising approach: helping instructors shape student–AI interactions from the very first click.