From Manager to Mentor: A Practical Strategy for AI Development
As faculty, we know that working effectively with our Assistant Instructors (AIs) is key to a successful course. In last week’s post on “Best Practices for Working with Assistant Instructors,” I highlight the importance of mentorship and creating professional development opportunities. But what does that mentorship look like in practice?
One of the most impactful ways to mentor our AIs is to equip them with high-leverage teaching strategies. Instead of just managing their grading, we can teach them how to teach. A powerful approach for this is the “Guide on the Side“ philosophy, which shifts the AI’s role from a simple answer-key to a learning coach.
The Guide on the Side: Coaching STEM Students in Problem-Solving
It’s a familiar scene in any STEM lab or office hour: a student, staring at a screen, is utterly stuck. For new teaching assistants (Associate Instructors, or AIs), the temptation is strong to take the shortcut; to grab the keyboard, write the line of code, or simply provide the answer. But while this solves the immediate problem, it bypasses a crucial learning opportunity.
This is where the Guide on the Side approach comes in. It’s a teaching philosophy that equips new AIs with practical strategies to coach students through the problem-solving process rather than solving problems for them. For faculty in STEM, empowering your AIs with these skills can transform your students’ learning experience.
Why This Shift in Pedagogy Matters
Across STEM disciplines, students frequently encounter “sticking points” moments of cognitive friction where the path forward isn’t obvious. If an instructor or AI simply hands over the solution, the student leaves with a single answer but no transferable skill. They learn to be dependent on an external expert.
By contrast, an instructor who guides the process models resilience, inquiry, and expert reasoning. The student leaves not only with a solution but with strategies they can apply to the next problem, and the one after that. They learn how to think.
Putting Theory into Practice: Activities for Your AIs
Faculty can use these activities in their own training sessions to help AIs develop a coaching mindset:
“Sticking Point” Brainstorm: In a think-pair-share format, AIs identify the most common places their students struggle. This builds a shared awareness of teaching challenges and normalizes the experience.
Scenario Analysis: AIs compare two contrasting dialogues: one where the AI gives the answer directly, and another where the AI uses Socratic questioning to lead the student to their own solution.
Questioning Roleplay: In pairs, AIs practice how to respond with guiding questions when students make common statements like, “I’m totally lost,” or “Can you just tell me if this is right?”
A Simple Framework for Modeling Expertise
A core strategy of this approach is teaching AIs to make their thinking visible. Experienced problem-solvers naturally follow steps that are often invisible to novices. Encourage your AIs to narrate their own problem-solving process explicitly using a simple four-step framework:
Understand: Restate the problem in your own words. What are the inputs, the desired outputs, and the constraints?
Plan: Outline possible approaches. What tools, algorithms, or libraries might be useful? What are the potential pitfalls of each approach?
Do: Execute the plan step by step, narrating the reasoning behind each action. (“First, I’m going to create a variable to hold the total because I know I’ll need to update it in a loop.”)
Reflect: Test the solution. Does it work for edge cases? Could it be more efficient? Are there alternative ways to solve it?
This explicit modeling teaches students how to think, not just what to do.
The Power of a Good Question: Building a Question Bank
Guiding questions are the primary tool of a “Guide on the Side.” They skillfully shift the cognitive work back to the student. Encourage your AIs to build a bank of go-to questions, such as:
To start a conversation: “What have you tried so far?” or “Can you walk me through your current approach?”
To prompt a next step: “What does that error message suggest?” or “What’s the very next small step you could take?”
To encourage deeper thinking: “Why did you choose that particular method?” or “What are the trade-offs of doing it that way?”
To promote reflection and independence: “How could you check your answer?” or “What would you do if you encountered a similar problem next week?”
Navigating Common Classroom Challenges
This approach provides concrete strategies for these common moments:
When a student is silent: Allow for sufficient wait time. If the silence persists, break the problem down and ask a simpler, first-step question.
When a student is frustrated: Acknowledge their feelings (“I can see this is frustrating; these problems are tough.”) and normalize the struggle before gently re-engaging with the task.
When a student just wants confirmation: Instead of giving a simple “yes” or “no,” redirect with a metacognitive prompt like, “What makes you confident in that answer?” or “How could you design a test to verify that?”
Resources for a Deeper Dive
For faculty and AIs who want to explore this pedagogical approach further, these resources are short, impactful, and highly relevant:
Book: Small Teaching: Everyday Lessons from the Science of Learning by James M. Lang
Article: Asking Questions to Improve Learning – Washington University in St. Louis Center for Teaching and Learning
Video: Eric Mazur’s video on Peer Instruction is a great resource for understanding how to shift from traditional lecturing to more active, student-centered learning. He effectively demonstrates the curse of knowledge and how students learning from each other can be more effective than an expert trying to explain something they’ve long ago mastered.
His approach, where students first think individually, then discuss with peers, and finally re-evaluate their understanding, directly aligns with the principles of guiding students through problem-solving rather than just showing them the answer. It emphasizes active processing and peer teaching, which are crucial for deeper learning and developing independent problem-solvers.
The Takeaway for Faculty
The “Guide on the Side” approach aligns perfectly with evidence-based teaching practices. By encouraging your AIs to slow down, model your thinking, and use questions effectively, you help them grow from being answer keys into becoming true teaching coaches. The result is a more engaged and resilient cohort of students who leave your courses not only with solutions, but with the confidence and strategies to tackle the next challenge independently.