Executive Snapshot
This self-initiated portfolio prototype explores how manager training can move beyond communication tips into judgment under ambiguity. I designed an interactive branching simulation that helps managers notice early performance risk across multiple 1:1 conversations before the problem becomes visible in delivery.
The prototype was built to demonstrate how I approach performance diagnosis, consequence-driven learning, branching logic, and AI-supported reflection in Articulate Storyline. It was not deployed in a live organization. I created it as a portfolio case study to show how I think through subtle people-performance problems where there is no single correct answer.
Business Context & Performance Problem
In many organizations, managers are expected to notice emerging performance risk early. In practice, they often do not miss the issue because they do not care. They miss it because:
the employee still sounds mostly functional
the signals are partial, not explicit
reassurance feels reasonable in the moment
escalation feels too heavy too early
The performance problem is not a lack of people-management theory. It is misjudgment across repeated ambiguous moments.
This prototype reframes manager development as practice in:
recognizing weak signals over time
distinguishing normal fluctuation from emerging pattern
choosing proportionate support
avoiding both premature escalation and passive reassurance
Experience the solution
Four simulated 1:1 conversations spread across time
Variable-driven consequence logic across the full scenario
Multiple endings based on the learner’s decision pattern
AI-supported reflection checkpoints that slow judgment down
AI-supported debrief at the end to help the learner extract a lesson and retry
Action Mapping (Diagnostic Backbone)
Although this was a portfolio prototype, it was designed using a full Action Mapping mindset, just like the CyberSecurity project structure on my portfolio.
Business Goal
Help managers detect emerging performance risk earlier, so support can stay proportionate and costly late-stage recovery can be reduced.
Critical On-the-Job Behaviors
Managers must be able to:
notice recurring signals across conversations, not just isolated comments
probe for usable detail without overreacting
distinguish broad reassurance from real clarity
choose when to support, wait, or escalate
interpret timing as part of the problem, not just the content of one conversation
Key Barriers Identified
Signals stay socially acceptable for too long.
Early strain often sounds like normal workload pressure.
Managers over-trust confidence language.
“Under control” can delay deeper inquiry.
Escalation feels too heavy until it is suddenly necessary. That compresses the recovery window.
The skill is interpretive, not procedural.
This is about judgment, not script memorization.
What Was Intentionally Excluded
generic active-listening theory
checklist-style manager tips
one-best-answer dialogue coaching
performance-management policy instruction
The focus is judgment under uncertainty, not concept recall.
Constraints & Friction
Even as a prototype, I designed it to reflect real managerial constraints:
conversations are short
information is incomplete
employees do not always disclose clearly, even when asked good questions
the same decision can help or harm depending on timing
consequences are often delayed, not immediate
These constraints shaped the narrative pacing, the variable logic, and the decision design.
Options Considered & Rejected
Several more conventional approaches were deliberately rejected.
1. Single branching conversation
Rejected because: it would reduce the problem to one conversational moment, when the real skill is noticing a pattern over time.
2. Linear case study with commentary
Rejected because: it could explain the issue, but not let the learner experience the consequences of their own interpretation.
3. AI-driven open conversation as the main simulation
Rejected because: it would reduce control over pacing, consequence logic, and replay consistency.
Instead, I chose a multi-week branching structure with controlled AI reflection, so the experience stayed realistic, replayable, and instructionally stable.
Design Strategy & Key Decisions
Key design decisions included:
1. Designing the experience around four time-separated 1:1s
→ This made the core skill pattern recognition over time, not isolated response selection.
2. Using variable-driven endings instead of one correct branch
→ This allowed the outcome to reflect the learner’s overall judgment pattern, not one lucky choice.
3. Making openness conditional
→ Good questions did not always produce immediate clarity. This preserved realism and prevented the simulation from feeling gameable.
4. Separating simulation from AI reflection
→ AI was used where interpretive support adds value: reflection and debrief, not primary scenario execution.
5. Building in replay as part of the learning loop
→ The learner is meant to test a hypothesis, see consequences, reflect, and retry for a stronger ending.
Learning Science
The prototype applies evidence-based principles:
Action Mapping: each learner choice maps to a real managerial behavior
Consequence-based learning: outcomes emerge from cumulative judgment, not isolated correctness
Contextual practice: learners work inside a credible managerial situation, not decontextualized instruction
Reflection for transfer: the learner pauses to interpret what happened and what to try differently next time
Replay value: multiple endings support iterative improvement rather than one-and-done completion
Context, Challenge, Gamification, Consequences
Context
The learner plays a manager in a recurring 1:1 cycle with an employee whose performance risk is not explicit at first. Across Weeks 1, 4, 7, and 10, the manager must interpret whether the situation is routine fluctuation or an emerging pattern.
Challenge
The challenge is not simply to “ask better questions.” It is to decide:
when to probe
when reassurance becomes avoidance
when support is proportionate
when waiting increases risk
when escalation is appropriate, and when it is premature
Consequences
All decisions culminate in one of five outcomes:
Great – early detection and light recovery
Good – pattern noticed, but later in the cycle
Neutral – mixed judgment, manageable but inefficient recovery
Bad – delayed clarity leads to heavier intervention
Worst – low signal detection and reduced openness narrow recovery options
The emotional and operational contrast between endings reinforces a core lesson: managerial risk often becomes expensive gradually, not all at once.
Implementation & Delivery
The prototype was built in Articulate Storyline with:
variable-driven branching logic
multi-ending consequence structure
controlled AI mentor reflections using JavaScript and the OpenAI API
reaction layers and transition slides to pace narrative interpretation
replay-oriented ending and debrief flow
What This Project Demonstrates About Me
I design learning for judgment under uncertainty
I make my decision-making and tradeoffs explicit
I can translate a subtle people-performance problem into a concrete simulation architecture
I know when AI adds value and when it reduces instructional control
I think beyond “content creation” and design for performance, consequences, and replay-based learning