Evaluating subjective, creative work publicly is one of the most cognitively and emotionally demanding assessment contexts. Under time pressure and emotional stress, reviewers often fall back on unstructured feedback. They either avoid difficult conversations, deliver vague critique, or unintentionally trigger defensive reactions.

The Business Impact:

• Lower Learner Retention: Disengaged students or employees due to poorly delivered critique.
• Increased Operational Costs: High volume of grade appeals, rework, and conflict resolution.
• Reputational Damage: 70% of feedback sessions fail to maintain psychological safety (HBR), leading to a loss of trust in the institution's assessment process.

The goal wasn't just to teach reviewers to “be nice”, but to be constructive, fair, and rigorous without triggering cognitive shutdown.

To bridge the gap between knowing the rules and applying them under pressure, I designed a Scenario-Based Learning experience. Learners step into the shoes of a new instructor conducting their first high-stakes review panel.

Instead of a passive information dump, the module throws learners directly into realistic, tense social dynamics.

Key Features:

• Immersive Social Dynamics: Scenarios mimic real-world friction (e.g., a distracted panelist, an awkward silence).
• Immediate Consequences: Choices instantly impact the student's visible anxiety level and the room's tension.
• Structured Reflective Framework: Learners practice a repeatable pattern: Strength → Clarify → Improve → Encourage.

Built with: Figma, Articulate Storyline, ChatGPT, Gemini Pro.

To ensure the simulation felt authentic, I mapped the target audience into two primary behavioral profiles. Understanding their distinct pain points allowed me to design realistic distractors (incorrect choices) that mirrored their natural instincts.

Knowing the distinct habits of our personas (the “Anxious Rookie” vs. the “Blunt Expert”), it was clear that a traditional information dump wouldn't change their behavior. Under stress, people default to their instincts. Giving them a list of “feedback rules” to memorize wouldn't help them when a colleague starts scrolling on their phone or a student becomes defensive.

Therefore, I shifted the instructional focus from Passive Recall to Real-World Decision Making.

Instead of asking learners what the rules are, the simulation tests how they apply them under pressure. I mapped the entire experience to four core behavioral objectives.

By the end of this simulation, reviewers must be able to:

• Regulate the room: Set clear expectations at the opening to reduce student anxiety and protect the schedule.
• Protect attention: Safely de-escalate panel distractions (e.g., phone scrolling) without shaming colleagues.
• Navigate awkwardness: Steer past uncomfortable silences or risky, biased comments professionally.
• Deliver balanced critique: Apply a structured framework (Strength → Clarify → Improve → Encourage) to prioritize actionable feedback under a strict time limit.

To deliver this project at scale without compromising on quality, I integrated a multimodal AI workflow into the design process. This allowed for rapid iteration and a 40% reduction in the initial production cycle.

• Narrative & Logic Design (LLMs): Used advanced prompting to brainstorm realistic workplace friction and draft complex branching dialogue. This ensured the scenarios felt authentic and psychologically grounded while saving weeks of manual scripting.
• Rapid Visual Prototyping (Gemini Pro): Leveraged AI to generate environment concepts and character references. This allowed me to establish the high-fidelity "comic-book" aesthetic and visual mood board in hours rather than days.
• Efficiency Metric: By automating the rote parts of content drafting and visual conceptualization, I reduced the design-to-prototype phase by ~40%, allowing more time for deep instructional alignment and UX testing.

The solution was delivered as a scenario-based eLearning simulation designed in Figma and developed in Articulate Storyline. Leveraging an AI-accelerated production pipeline, the prototype uses immediate visual consequences to simulate the tension of a live panel. Instead of traditional quizzes, the experience focuses on qualitative behavioral choices and explanatory feedback. The module is SCORM-compliant and ready for deployment via any LMS or as a standalone web experience.

Along with the interactive simulation, I designed a mobile-friendly Job Aid. I included this as a performance support tool so learners don't have to memorize the entire module — they can simply pull up the image before walking into a real review session.

If I were to iterate on this project, I would integrate a conversational AI role-play component (using an LLM API) for the final assessment. This would allow learners to type their feedback rather than selecting multiple-choice options, dramatically increasing the fidelity of the practice.

This case demonstrates that behavior change in high-stakes environments cannot be achieved through information recall alone: under stress, people default to their instincts.

By shifting from passive instruction to immersive, consequence-driven scenarios, I created a safe space to practice navigating social friction. This approach bridges the gap between theory and real-world application, equipping reviewers to deliver actionable critique while actively protecting psychological safety.