Introducing Chat mode: AI Conversations
ImmerseMe brings real-world language into the classroom through authentically filmed 360° scenarios. Students can explore a place, hear language in context and build confidence through structured, scripted practice.
Now, with Chat mode, they can take that practice into a dynamic, unscripted conversation.
Students have their own questions to ask, choices to make and ideas to express. Their AI conversation partner responds to what they say, opening up different directions within a familiar situation. Two students starting the same lesson can have very different conversations.
For me, this is an exciting step forward: combining the grounding of our filmed experiences with the freedom to communicate spontaneously, supported by clear learning objectives and performance and proficiency rubrics.
Familiar situations with new possibilities.
Our initial Chat lessons draw their situations and learning objectives from ImmerseMe's existing 360° scenarios. Each combines:
A situation.
An AI conversation partner.
Up to six learning objectives.
A desired proficiency level.
The objectives give the conversation an academic purpose while leaving students room to decide how they approach it.
In one of our team's Spanish tests, the learner visited a food stall. They asked for recommendations, chose a meal, added a drink and discussed payment. The vendor responded to their requests and asked follow-up questions.
That creates opportunities to draw on familiar language, respond to something unexpected and find another way to express an idea. Beginners might ask simple questions and express preferences; more advanced learners might explain choices and explore different viewpoints.
Teachers can connect these experiences with current units, using structured practice to prepare students before they apply their language more independently in Chat.
Understanding performance.
Students earn XP and receive a Chat score out of 100, measured against the lesson's desired proficiency level. An 85 in a beginner lesson reflects different expectations from an 85 in an advanced lesson.
The score combines four areas:
Content (30%): Achieving the objectives and contributing relevant, developed ideas.
Fluidity (30%): Connecting sentences and ideas through varied structures and smooth transitions.
Vocabulary (20%): Using words and expressions with appropriate range, relevance and precision.
Grammar (20%): Controlling the target language's structures, considering the frequency and impact of errors.
Half of Content comes from objective completion and half from contribution quality. Students receive credit for achieving the task while still having room to strengthen how they communicate.
Recognizing demonstrated proficiency.
A separate placement rubric considers the proficiency demonstrated across the conversation.
It looks across ImmerseMe's levels, displayed through frameworks such as ACTFL, from Novice Low to Distinguished, or CEFR, from A1 to C2. Each category requires consistent evidence across the exchange, so one impressive sentence or difficult reply should not determine the result.
This can reveal strengths and areas for development, such as stronger vocabulary than grammar. It also recognizes that a short, simple task may offer limited opportunities to demonstrate advanced language. Placements support teacher judgment rather than provide formal proficiency certification.
Respecting how each language works.
The rubrics include guidance for assessing languages on their own terms. Spanish can leave a subject unstated, Japanese can omit information understood from context, and Mandarin expresses time differently from languages with extensive verb conjugation.
The assessment considers the structures each language actually uses. Pronunciation feedback on individual spoken replies remains separate from the overall conversation score.
Giving teachers insight into independent communication.
Teachers can review transcripts alongside objectives and scores to see how students approach an unscripted exchange. Did they initiate a question, develop an answer or respond to a suggestion?
Students can revisit an exchange and improve a response, while teachers use patterns across conversations to guide follow-up teaching.
We've also expanded our base-language options, helping more learners access the interface and translations. Chat Rules reinforce privacy, respectful behavior and awareness that AI can make mistakes, while moderation helps identify content needing review.
Building toward greater customization.
Today's Chat lessons build on our existing scenarios. Looking ahead, we plan to give teachers the ability to define their own learning objectives, connecting conversations even more closely with their curriculum.
I'm excited to see students take the language they've practiced and make it their own. Try a familiar scenario in Chat mode and see where the conversation goes!
For practical classroom recommendations and help interpreting results, read our Help Center guide for Using Chat mode in your classroom.
Ngā mihi nui,
Scott Cardwell
Co-founder and CEO, ImmerseMe
