The AI Conversation Simulator
A student who has earned C’s all year turns in a flawless essay. A parent writes, angry, after you asked their child about a submission. The AI Conversation Simulator puts educators in those moments, lets them write the message they would actually send, and shows how it lands before a real student or parent reads one like it.
The case
Most educators will have this conversation, and few have practiced it. The message that opens it decides whether a student explains what happened or goes quiet.
Starting with “I noticed” and a question about process keeps the student talking. An accusation ends the conversation, even when the suspicion is right. Whatever the student did, the opening message decides whether the teacher hears the real story or a defense.
AI detection tools produce false positives and cannot show how a piece of writing came to be. A conversation built on a detector score starts with a claim the teacher cannot support, and students and parents know it.
The coaching stance separates the work from the person, asks about process instead of judging the product, and always leaves a path forward.
Reading about the right approach is not the same as writing the message under pressure. The simulator gives feedback on the actual words, shows the reply they earn, and lets the educator revise and try again.
The tool
The AI Conversation Simulator is a browser-based practice tool built on two ideas from Inclusive AI Strategies workshops: the difference between a coaching stance and a catching stance, and a graduated response (notice, invite, respond, involve) that keeps escalation as a last step rather than a first one.
Four tracks match the participant’s teaching context: writing-intensive courses, STEM, electives, and world languages. Each holds a set of scenarios: a student whose writing changed overnight, a student who admits to falling behind, a portfolio with no process behind it, a personal reflection with no one in it, a parent who wants a straight answer, a class where the pattern is spreading. The world languages track adds the cases that only come up when the subject is the language itself: a paragraph written three chapters ahead of the syllabus, a heritage speaker flagged for sounding like one, a speaking recording with no hesitation in it, and a class that cannot tell a dictionary from a translator. Some scenarios ask for a written message; others present choices and play out the consequences.
In a written scenario the loop runs like a real exchange. The participant drafts the message and gets feedback on its coaching and catching language, with the exact phrases quoted and a replacement offered. The student or parent then replies in character to what was actually written. The participant writes one more message, gets feedback on it in context, and sees how the exchange landed: opening up, guarded, or shutting down, with a short note on why. Every written scenario also shows a model response in the coaching stance, and the session ends with a toolkit the participant can print or copy.
The feedback and the replies come from an LLM through an API, which does not use the text to train the models. A fixed framework the participant cannot change guides the feedback, and the character who replies is a learner who did use AI and is anxious about being caught. The exchange may be accessed by Inclusive AI Strategies to make future improvements to the simulator. The privacy policy has the details.
In the room
The simulator is part of Inclusive AI Strategies keynotes and workshops on AI and academic integrity, including What Do I Actually Say? The facilitator puts an access key on the screen, or a QR code that contains it, and participants open the simulator on their own phones or laptops. No accounts, no installs.
The key stays valid for a week after the session, so participants can come back to the scenarios they skipped or try a different track. A facilitator view lays out every scenario, option, and consequence on one page for the debrief, and a usage log shows how the room engaged with each scenario without storing anyone’s name.
Getting started
Your facilitator shares the key during the session, on the screen or as a QR code.
Inclusive AI Strategies brings the simulator to conferences, districts, and campuses as part of a keynote or workshop on AI and academic integrity.
Inclusive AI Strategies