Roleplay + Feedback Coach
An AI-native learning experience prototype
An AI-native learning experience prototype
This is a browser-based microlearning tool that puts learners inside a real conversation and coaches them on the spot.
The learner is dropped into a scenario — a tense family member in a hospital hallway, a defensive direct report, a skeptical VP — and has to respond in writing, in real time, to an AI persona that reacts authentically to whatever they say. If they lead with empathy, the persona softens. If they're vague or dismissive, the situation escalates. The feedback loop is immediate and consequential in a way that a multiple-choice question never is.
After two or more exchanges, the learner can request coaching feedback at any time. The feedback engine analyzes the full conversation transcript and returns a score out of 10, a set of strength tags, a set of growth areas, a specific coach's note tied to what the learner actually wrote, and — when relevant — a stronger alternative phrasing they could have used. Everything is generated fresh against the real transcript, so no two feedback reports are identical.
Scenarios included:
-Nurse delivering news to an upset family member
-Manager giving constructive feedback to an underperforming direct report
-L&D specialist pitching a new program to a skeptical executive
Each scenario can be swapped or extended without touching the core application logic — the scenario definitions are a single configuration object.
What makes it differentMost eLearning simulations are branching scenarios with pre-scripted paths. This one has no script. The AI persona improvises within a defined character and emotional state, which means learners can't game it by clicking until they find the right answer. They have to actually communicate.
The feedback is similarly unscripted. Rather than a rubric with preset responses, the coaching engine reads what the learner wrote and responds to it directly — which means a learner who opens with "I know last week was rough, can we talk about what got in the way?" gets different feedback than one who opens with "We need to talk about your performance."