Judgment + simulation

Steam Support

Three realistic customer emails become practice in diagnosing incomplete evidence, choosing a move, and writing a response designed to earn trust.

FocusCustomer judgment
StructureDiagnose → decide → respond
FeedbackLikely customer reaction
BuildLive simulation
What this demonstrates: customer-support performance depends on interpreting incomplete evidence and choosing a defensible response. Policy recall alone is not enough.

Context

Support work is a judgment problem disguised as an information problem.

Customer-service professionals rarely receive a perfectly structured case. They receive partial evidence, emotion, conflicting signals, and a customer who expects a useful answer. The challenge is not simply knowing the policy. It is deciding what the situation means and what to do next.

Problem

Traditional support training often separates knowledge from the moment of use. Learners study policies, examples, and process steps, then encounter a real customer whose situation does not match the clean examples. The gap is judgment under uncertainty.

Decision

Make the learner diagnose before they answer.

The simulation is organized around three support emails covering technical issues, billing, and account access. Each case contains incomplete evidence. The learner must determine what is happening, assess the risk of different moves, choose a response strategy, and write or assemble a customer-facing reply.

Feedback

Feedback is framed around consequence rather than correctness. Instead of only saying that an answer is right or wrong, the experience surfaces the likely customer reaction and the risk created by the response. That makes the reasoning behind the decision visible.

Design principle: realistic practice should expose the tradeoff. If every option except one is obviously ridiculous, the learner is not practicing judgment.

My role

I designed the simulation structure, ticket scenarios, response-option risk model, and feedback writing. The project demonstrates the combination of learning strategy, scenario writing, interaction design, and hands-on web production.

Evidence

This is an independent portfolio concept, so there are no invented completion, satisfaction, or Level 4 results. The functioning build is the evidence: three distinct cases, a repeatable decision model, realistic customer language, and consequence-based feedback.

Reflection

The learning target was never “write a nice email.”

The target was to form a better judgment about what the customer needs, what the available evidence supports, and how much confidence the representative should place in a proposed action. The writing is the visible output of that reasoning.

Independent portfolio concept. Not affiliated with, sponsored by, or endorsed by Valve or Steam. The concept does not disclose proprietary Valve or Steam information.